1 ABSTRACT This thesis investigates existing research on the linguistic evolution of Latin, Spanish, and French in order to form an in-depth, multilingual analysis of the development of the subjunctive mood in Romance over time. The diachronic nature of the conclusions of this research are supported by synchronic examinations of individual languages. Six distinct subjunctive systems are studied within the thesis: Classical Latin, Vulgar Latin, Old Spanish, Modern Spanish, Old French, and Modern French. Each subjunctive system is evaluated through the dual lenses of morphology and pragmatics. That is, within every language, changes in both grammatical mood formation as well as semantic use of the subjunctive are evaluated. Elements of stability and variability are identified in each language, and observations made during this investigation are connected to established theories of general language change. Finally, predictions for the future of the subjunctive mood in Romance are made using evidence from the research. Overall, this thesis finds that while superficial changes in the subjunctive moods of Latin, Spanish, and French abound, the core significance and usage of the subjunctive remains relatively unchanged. 2 Through the Looking Glass: the Development of the Subjunctive Mood in Romance Honors Thesis Proposal by Elizabeth Belka Committee members: Esther Castro (Chair) Prof. Justin Crumbaugh Prof. Nieves Romero-Diaz Prof. Marta Sabariego Prof. Carolyn Shread 3 ACKNOWLEDGEMENTS First and foremost, I dedicate this thesis to my mother, the incredible Angel, who carefully and tirelessly read and reread every word I wrote, despite having no background in linguistics or language studies. I also dedicate this thesis to my amazing father, George, who among other support lent me his desk and laptop setup for the entirety of spring break, as well as my twin sister, Victoria, who made time to support me even in the midst of her own thesis research and is generally my greatest inspiration in life. I would also like to acknowledge my entire extended family for cheering me on from afar, and for sending me love when I needed it most. This thesis would not have become a reality without my faculty advisor, Prof. Esther Castro, to whom I am greatly indebted for the sheer amount of patience she has shown me for the past year (and indeed, my entire undergraduate career). I fear I may have presented more of a challenge than she originally anticipated, but she never once gave up on me or my ideas. My extraordinary thesis committee-- Profs. Crumbaugh, Romero-Diaz, Sabariego, and Shread-- also deserves my utmost gratitude for their unwavering support. Although they were not privy to the behind-the-scenes chaos, I would also like to thank my high school students at South Hadley High School, for bearing with me as their student teacher during the busiest semester of my life. I thank the courses of Spanish 5 and Honors Spanish 4 in particular for being the first classes to whom I taught the subjunctive mood. My experience in the classroom has deepened my understanding and appreciation of explaining grammatical concepts to those politely feigning interest. 4 Another pillar of support I found in South Hadley was my Supervising Practitioner, Paula Lonergan, who graciously allowed me into her classroom and did everything she could to help me as an educator, a researcher, and a human being. Plūrimās gratiās vōbīs agō, muchísimas gracias, merci beaucoup. 5 TABLE OF CONTENTS ABSTRACT 1 ACKNOWLEDGEMENTS 3 TABLE OF CONTENTS 5 LIST OF TABLES AND FIGURES 6 1. INTRODUCTION 7 2. THE MORPHOLOGICAL EVOLUTION OF THE SUBJUNCTIVE MOOD: FROM CLASSICAL LATIN TO MODERN ROMANCE 25 3. THE PRAGMATIC EVOLUTION OF THE SUBJUNCTIVE MOOD: FROM CLASSICAL LATIN TO MODERN ROMANCE 67 4. THE SUBJUNCTIVE IN MODERN ROMANCE AND SECOND-LANGUAGE LEARNERS 106 5. THE FUTURE OF THE SUBJUNCTIVE 122 BIBLIOGRAPHY 130 6 LIST OF TABLES AND FIGURES Table 1: The Active Subjunctive Paradigm in Classical Latin 11 Table 2: The Active Person Endings in Classical Latin 27 Table 3: The Active Subjunctive Paradigm in Vulgar Latin 27 Table 4: The Active Subjunctive Paradigm in Old Spanish 39 Table 5: The Active Subjunctive Paradigm in Modern Spanish 42 Table 6: The Active Subjunctive Paradigm in Old French 46 Table 7: The Active Subjunctive Paradigm in Modern French 50 Table 8: Coinciding Morphology of the Latin Perfect Subjunctive & Future Perfect Indicative 59 Table 9: Uses of the Subjunctive in Classical Latin Independent Clauses 67 Table 10: Uses of the Subjunctive in Classical Latin Dependent Clauses 69 Table 11: Latin Protactic to Hypotactic Shift 75 Table 12: Uses of the Subjunctive in Modern Spanish Independent Clauses 83 Table 13: Uses of the Subjunctive in Modern Spanish Dependent Clauses 85 Table 14: Uses of the Subjunctive in Modern French Independent Clauses 93 Table 15: Uses of the Subjunctive in Modern French Dependent Clauses 94 Figure 1: The Pragmatic Evolution of the Romance Subjunctive 99 Figure 2: The Syntactic Evolution of the Romance Subjunctive 100 7 “In another moment Alice was through the glass, and had jumped lightly down into the Looking-glass room…” -Through the Looking Glass by Lewis Carroll 1. INTRODUCTION When imagining the very beginnings of human language, one might picture words used as a communicative tool to describe the physical presence of objects within a general space, a caveman pointing at an open flame and arbitrarily producing sounds that thousands of years later will become the word fire. Although such labels and descriptions are an important function of language, at times it is necessary even for the caveman to express not only what is or was “there”, but indeed what could or should be “there”. In the study of linguistics, this type of expression belongs to the semantic category of modality. The linguistic function of modality exists in order to discuss situations that are possible. As a category, modality can be broken down further, notes Murphy (2008) in The role of typological drift in the development of the Romance subjunctive: Modality is, first and foremost, a semantic category that “covers a broad range of semantic nuances—jussive, desiderative, intentive, hypothetical, potential, necessitative, inferential, hortatory, exclamative, etc.—whose common denominator is the addition of an overlay of meaning to the most neutral semantic value of the proposition of an utterance, namely factual and declarative” (Fleischman 1982: 13). Modality can also be divided into two major types: epistemic (knowledge-based) and deontic (action-based). Epistemic modality involves notions of speaker commitment, opinion, judgment and 8 evidence, while deontic modality represents either internal or external need or compulsion (Bhat 1999: 63). (p. 2) Thus, referring back to the earlier example of a caveman, an utterance with “the most neutral semantic value of the proposition” would be a factual statement like this is fire. However, consider the difference in connotation if our caveman were instead to say something like this should be fire or this may be fire. In the second two phrases, the physical existence of fire is put into doubt, and the perceived context of the situation alters. Rather than picturing a caveman gleefully exclaiming the birth of a flame, we may conjure the image of the caveman fruitlessly rubbing sticks together or looking speculatively at a column of smoke. A relatively small semantic change has led to a large adjustment in pragmatic significance. 1.1.1 Grammatical Mood This “overlay of meaning” of modality can be grammatically expressed in different ways, the most prevalent of which in the Romance languages is mood. Mood depends on verbal inflections (changing the actual ending of the verb) to express modality, as opposed to other expressions, which rely on auxiliary verbs or clitics (adding additional words like “must” or “should” to the verb). The Romance languages today have four distinct moods: the conditional, the imperative, the indicative, and the subjunctive. Of these moods, only the indicative is considered a “realis” mood, in the sense that the indicative is used to express facts or known events in the “real” world. The verb in a “realis” mood indicates that, from the point of view of the speaker, an eventuality that is, was, or will be “real”. An example of a verb in the indicative mood in English is the “is” from the sentence “there is a fire”. The speaker is pointing out a physical phenomenon in the material world shared with the listener; the fire exists concretely as an eventuality. 9 The three remaining moods are all considered “irrealis” moods, and indicate that, to the knowledge of the speaker, an eventuality is “unreal”. The conditional mood is used with verbs whose fulfillment depends on a condition, as in the sentence “if the sticks were dry, there would be a fire by now”. In this sentence, the fire does not exist; the speaker is expressing that one could have existed at the time given a condition that was unfulfilled (the sticks being dry). The eventuality of the fire existing in the speaker’s timeframe is “unreal”. The imperative mood expresses direct commands, like “light the fire!” Similarly to the previous example sentence, in this case the fire does not exist. The speaker is demanding that someone else create the fire, so the eventuality of the fire existing in the speaker’s timeframe is “unreal”. Grammatical mood should not be confused with grammatical tense: sentences like “there is a fire”, “there has been a fire”, “there was a fire”, “there had been a fire”, and “there will be a fire” are all in the realis mood of the indicative. The eventuality of the fire existing in each of these sentences is “real”. As previously explained, grammatical mood can also express “unreal” eventualities, like in the sentences “there could be a fire” or “put out the fire”. Grammatical tense, on the other hand, differentiates the time frame of a verb eventuality, like the past or present tenses express eventualities in the present or past time frame of the speaker’s point of view, respectively. Therefore, while the verbs in the sentences “there is a fire” and “there was a fire” both belong to the same grammatical mood (the indicative), they belong to different grammatical tenses; “is” is the present tense referring to an eventuality taking place in the present, and “was” is the past tense referring to an eventuality taking place in the past. 10 Grammatical moods like the indicative can support many tenses, while others, like the conditional, only contain one tense. 1.1.2 The Subjunctive Mood The third and last of the irrealis moods in the Romance languages is the subjunctive mood. Native speakers and language learners alike may encounter difficulties when using or explaining the subjunctive due to its wide array of applications; English speakers in particular may struggle with fully understanding the mood because in many situations there is no direct English equivalent. The subjunctive appears in subordinate clauses in sentences to express opinions or emotions, hypothetical events, wishes or desires, or polite requests, among other more esoteric employments. Comprehension of the subjunctive mood usually relies on its contrast with the indicative. For example, notice the distinction in the sentences “I was an athlete, so I made the goal” and “If I were an athlete, I would have made the goal”. The first sentence uses the indicative form of the verb to be, i.e. “I was”. However, the second sentence uses the subjunctive form. In English, saying “I were an athlete” is grammatically incorrect unless it is preceded by the “if”, which changes the modality of the clause. “I was an athlete” expresses a factual statement that the encoder was indeed an athlete in the past; the eventuality of being an athlete is “real”. On the other hand, “if I were an athlete” expresses the opposite; the encoder was not actually an athlete in the past, and the eventuality of being an athlete is “unreal”. It may be helpful to visualize this distinction using a metaphor: if the indicative mood depicts the “real world” where Alice begins her story in Alice in Wonderland, then the subjunctive mood depicts the world “through the looking glass” or Wonderland, where everything is almost the same, yet completely different. 11 Like the indicative mood, within the Romance languages system as well as in other language families, the subjunctive mood can support a collection of grammatical verb tenses. Often, a given language will be able to express the subjunctive in the present and in the past time frames. An additional nuance may exist within the grammatical tenses to distinguish verbal aspect. Aspect refers to an eventuality’s state of completion or incompletion within a certain time frame. Both Classical Latin and modern Spanish divide the “past” tense time frame of an action into either a “preterite” completed state or an “imperfect” incomplete state. To illustrate this contrast, consider the Spanish preterite “yo comí” and imperfect “yo comía” in examples (1a) and (1b). (1) a. Ayer yo comí la cena muy temprano porque no comí el almuerzo. Yesterday I ate the dinner very early because no ate the lunch. “Yesterday I ate dinner very early because I didn’t eat lunch.” b. Como niña, yo comía cada semana en este café. As girl. I ate every week in this café. “As a girl I ate every week in this café.” “As a girl I used to eat every week in this café.” Although both sentences can be translated into the English “I ate”, the preterite “yo comí” is only used when the speaker is referring to an action or state that is expressed as having a definite beginning and end point in the past; the eventuality is “complete”. In contrast, the imperfect “yo comía” is used for an event or state in the past that is not expressed as having a set start and/or end point (either because it is unknown or irrelevant for a communicative purpose) or that was repeated habitually, a loose equivalent of the English phrase “used to”. The imperfect signals an “incomplete” eventuality. “I ate” and “I used to eat” both occur in the past frame of reference, 12 but differ in terms of grammatical aspect. This aspectual dichotomy of “complete” and “incomplete” eventualities continues throughout the present and future tenses. 1.2 The Subjunctive Mood from Latin into Romance Although study of the subjunctive mood in Latin and the Romance languages began centuries ago, it continues to this day, likely due to the complex nature of the subject. Murphy (2008) comments that “[I]t becomes clear, even before conducting a closer examination of the literature, that opposing viewpoints are the norm rather than the exception with regard to the subjunctive mood in Romance languages” (p. 6). While the majority of scholars agree on specific paradigms of morphological forms at a set point in time in a given language, both diachronic and comparative synchronic research have proven to be polemic. Theories of origin and evolution have been written and rewritten as the modern Romance languages continue to change, and today’s predictions for the future of the subjunctive remain just as uncertain as they were 500 years ago. As is common in every discipline, scholars also disagree on whether or not the actual changes in grammatical structures and pragmatic usage can be “positive” or “negative” for society, as well as whether or not these changes are part of a larger pattern in a language family or unique to the individual circumstances of a language. 1.2.1 The Subjunctive Mood in Classical Latin Classical Latin, also known to some scholars as the “Golden Age”, is a period of written and spoken Latin usually ascribed to the period between 75 B.C.E. until the third century C.E. (Tara 2018). Classical Latin, still taught today as the literary standard of the language, featured a 13 morphologically robust and symmetrical verbal system. In the text Romance Languages: A Historical Introduction, Alkire and Rosen elaborate that The Latin verb system, itself the product of drastic innovation on the way from Indo-European, took on a squarish architecture characterized by three binary contrasts: voice (active and passive), mood (indicative and subjunctive), and aspect (infectum and perfectum), in addition to the familiar category of tense (present, past, and future). (127) In the active voice, Classical Latin developed four subjunctive tenses, each with its own tense and aspectual values (see Table 1). Each tense also corresponds to a tense in the active indicative verb system. The present and perfect categories share a present tense distinction, while present has an incomplete aspect and perfect has a completed one. In the same way, the imperfect and the pluperfect share a past tense distinction, with the imperfect having an incomplete aspect and the pluperfect having a completed aspect. Table 1: The Active Subjunctive Paradigm in Classical Latin A S P E C T Incomplete Complete T E N S E Present Present ego amem - I love tū amēs - you love is/ea/id amet - he/she/it loves nōs amēmus - we love vōs amētis - you all love eī/eae/ea eament - they love Perfect ego amāverim - I [have] loved tū amāverīs - you [have] loved id/ea/id amāverit - he/she/it [has] loved nōs amāverīmus - we [have] loved vōs amāverītis - you all [have] loved eī/eae/ea amāverint - they [have] loved Past Imperfect ego amārem - I loved Pluperfect ego amāvissem - I had loved 14 tū amārēs - you loved is/ea/id amāret - he/she/it loved nōs amārēmus - we loved vōs amārētis - you all loved eī/eae/ea amārent - they loved tū amāvissēs - you had loved is/ea/id amāvisset - he/she/it had loved nōs amāvissēmus - we had loved vōs amāvissētis - you all had loved eī/eae/ea amāvissent - they had loved Note: The regular inflections are bolded. Classical Latin made use of the subjunctives in several different ways, both in main and subordinate clauses. In main clauses, instances of the subjunctive can be broken down into three broad modal categories: optative, which indicates a hope or wish; volitive, which indicates will; and potential, which indicates opinion. The volitive itself can be further subdivided into five functions: the hortatory, which expresses a command in the first person plural; the jussive, which expresses a command in the third and occasionally second person; the prohibitive, which expresses prohibitions and negative commands; the concessive, which expresses something granted or conceded; and the deliberative, which expresses questions or exclamations. The potential category is also separated into uses based on English auxiliary verbs: the “may” potential conveys a mere possibility; the “would” an understood condition; the “could” perceiving or thinking; and the “should/would/could have” conveys irrealis. For a more in-depth discussion on these classifications and examples, please refer to Murphy (2008). In addition to extensive applications in main clauses, the subjunctive mood is also employed in Classical Latin as a marker of subordination. There are thirteen general scenarios in which the subjunctive appears, as enumerated in Murphy (2008): final (purpose) clauses expressing ‘in order that’ or ‘lest’; consecutive (result) clauses expressing ‘that [not]’ or ‘so 15 that’; “anticipatory” temporal clauses; protasis of “ideal” conditional clauses; protasis of “unreal” conditional clauses; comparison conditional clauses expressing ‘as if’; proviso clauses expressing ‘provided that’; adversative & Concessive clauses expressing ‘although’; clauses of characteristic; causal clauses; Oratio Obliqua (indirect discourse); circumstantial “cum” clauses; and indirect questions. Of course, the specific tense of the subjunctive used in each of these scenarios is determined by tense/aspect relationships within the sentence. 1.2.2 The Subjunctive Mood in Vulgar Latin Vulgar Latin, also known as Popular Latin, is directly related to Classical Latin in that while Classical Latin tends to refer to a written academic standard language, Vulgar Latin is the variety of the language spoken during the time of Classical Latin through its direct evolution into Late Latin. Despite existing concurrently, Vulgar and Classical Latin differed; as the “ideal” language of prestige among the elite, Classical Latin was codified and regulated, while Vulgar Latin varied depending on region and educational background of the speaker. As an export of the expansive Roman Empire, many speakers learned Vulgar Latin as a second or third language, and their language use was influenced by the linguistic profile of their region. Over time, Vulgar Latin increasingly diverged from Classical Latin until around the 6th century, when scholars classify it as having evolved into a completely distinct language (Tara 2018). Both the indicative and subjunctive verbal systems of Classical Latin underwent systematic changes as Vulgar Latin developed. Although each of the active subjunctive tenses (present, perfect, imperfect, and pluperfect) remained as categories, only the present remained fundamentally the same after accounting for regular sound changes and analogical adjustments. The original verbal endings of the imperfect subjunctive, however, disappeared completely. In 16 their place, new morphology for the imperfect was created by recycling endings from both the original pluperfect indicative and pluperfect subjunctive forms. At the same time, the original endings from the pluperfect and perfect subjunctive were gradually usurped by the periphrastic perfect system, in which an alternative perfect was introduced using the past participle in conjunction with the verb “habere”, or “to have”. While originally, the simple preterite tense could alternately express either perfective or perfect aspects depending on semantic context (e.g. the pragmatic meaning of both I sang and I have sung are articulated by the same Latin form ego cantavi), the increasingly popular periphrastic perfect construction eventually came to signify the true perfect, leaving the preterite to only express perfective aspect. This change is shown in examples (2a) and (2b). (2) a. Ego cantavi I sang “I sang” / “I have sung” b. Ego cantavi Ego habeo cantatu I sang AND I have sung “I sang” “I have sung” In the realm of pragmatics, Vulgar Latin saw a reduction of the usage of certain subjunctive forms, likely due to sound changes and syncopated or shortened verb forms which blurred the morphological tense distinctions of verb endings. In this period, the indicative began to fill in for the subjunctive in some subordinate clauses, and even in some main ones. 1.2.3 The Subjunctive Mood in Romance Languages Morphologically, the subjunctive system in the Romance languages more closely resembles that of Vulgar Latin, as opposed to Classical Latin. Linguistic evolution into the modern Romance 17 languages involved even more sound changes , which in turn affected the individual verbal endings. Nevertheless, as a whole, each of the main Romance languages have similar relationships with the subjunctive mood as inherited from Vulgar Latin. In all the modern languages, the present subjunctive tense is directly derivative of the present subjunctive seen in Vulgar Latin. In addition to the present, the modern Romance languages each developed and employed at one point the three remaining tense categories of perfect, imperfect, and pluperfect subjunctive. As in Vulgar Latin, the perfect and pluperfect subjunctive morphologies are formed using the descendent of “habere” with the past participle of the verb. Some languages, mainly Spanish and Portuguese, created a future subjunctive with morphological endings from the Latin future perfect indicative and perfect subjunctive, which had fallen into disuse. The pragmatic changes in the use of the subjunctive mood from Vulgar Latin to Romance parallel the shifts made between Classical and Vulgar Latin. Although the adaptation away from using subjunctive forms has been gradual over the centuries, in a cumulative review the results of this evolution may appear quite dramatic. After a systematic review of differences in mood in Latin and the Romance languages, Murphy (2008) concludes that “the verbs in approximately half of all subjunctive contexts in Latin have been replaced by non-subjunctive forms in Spanish and French” (p. 22). In certain scenarios, especially in main clauses, the use of the subjunctive has been relatively preserved, while in others, particularly subordinate clauses, the subjunctive is no longer favored. The corresponding indicative forms may replace the subjunctive in these cases, but increasingly popular is the employment of the conditional mood, which was developed in Vulgar Latin. 1.2.4 Evolutionary Changes 18 As long as scholars have noticed language change, there have been continual efforts to pinpoint the exact causes and reasons for it. The most coherent theory is that of linguistic equilibrium: language tends to seek balance, and when disrupted, changes in order to reestablish equilibrium. Begioni 2010 elaborates that Dans cette perspective les évolutions sur l’axe temporel correspondent aux passages successifs d’un système à un autre. La langue est en équilibre systémique à une époque t1, elle subit des changements linguistiques surtout au niveau de la morphologie et de la syntaxe qui ne sont, dans un premier temps, que des microvariations et qui au fur et à mesure deviennent des variations plus importantes. Celles-ci provoquent un déséquilibre du système qui est obligé de changer un certain nombre des règles de fonctionnement pour retrouver un nouvel équilibre à une époque t2. (pp. 14-15) In this perspective the evolutions on the temporal axis correspond to successive transitions from one system to another. Language is in systemic equilibrium in one time t1, it undergoes linguistic changes, especially at the morphological and syntactic level, which are, at first, only microvariations and gradually become more significant variations. These provoke a disequilibrium in the system, which is obliged to change a certain number of operating rules in order to establish a new equilibrium at a time t2. At first, these small “microvariations” caused by various social and psychophysical factors may go unnoticed or have no lasting effect on a language. However, some small shifts in morphology, phonology, or syntax may accumulate to such a level that systematic change occurs. Researchers also stress language’s predisposition to change; Aitchison (2001) theorizes that “there are often identifiable ‘weak spots’ in a language structure where change will be likely to strike, as well as stable elements which are likely to resist change” (p. 137). Many of these “weak spots” can be identified in the subjunctive system as early as Classical Latin, and with the evolution through to modern Romance languages, additional points of weakness develop and are exploited. Along with patterns of language change, patterns of restoring linguistic equilibrium can also be readily identified. Recovering system balance usually involves processes of reducing 19 complexity; “language minimizes opacity in that it lessens confusing ‘opaque’ situations, and maximizes transparency in that it moves towards constructions which are clear or ‘transparent’” (Aitchison, 2001, pp. 177-178). As expected, these same evolutionary patterns can be identified in the Romance languages. 1.3 Research Objectives The following thesis intends to conduct a diachronic examination of the subjunctive verbal system as it evolves both morphologically and pragmatically through Classical Latin into the modern Romance languages and beyond. While similar studies have been conducted, this thesis is unique in the depth and breadth of its examination, and meticulously yet clearly documents and evaluates the entire subjunctive systems in Classical Latin, Vulgar Latin, Old Spanish, Modern Spanish, Old French, and Modern French, not only in terms of verbal inflections, but also in terms of psychosocial significance of implementation. Particular focus will be placed on identifying and analyzing patterns of linguistic evolution and specific processes of language change, in hopes of contributing to the understanding of linguistic evolution and the study of why and how languages develop. The thesis enumerates first the subjunctive verb forms in Chapter Two, then their various social implementations in Chapter Three. Chapter Four uses an analysis of cross-linguistic shifts in conjunction with existing theories of language evolution, including the complex relationship between the subjunctive and second language learner, in order to predict further development of the Romance subjunctive. Chapter Five presents the conclusions of the thesis using evidence collected from all previous chapters, as well as recommending paths for further study. 20 Although a brief comparison of the subjunctive systems in Classical Latin, Vulgar Latin, Modern Spanish, and Modern French would point to the “loss” of certain synthetic verb forms and pragmatic uses of the subjunctive, to conclude that languages generally become simpler over time would ironically be an oversimplification of linguistic evolution. The processes of language development are at once easily understood and frustratingly complex, especially if the researcher retains a perspective that views change as a threat. This thesis endeavors to embrace the evolutions of the subjunctive past, present, and future, and argues that while the forms themselves may vary through the centuries, expressive power remains the same; in other words, the morphologies of the subjunctive are not constant, but the meaning of the mood itself is. As explained in the previous section, the most comprehensive theory of why languages change involves the cycle of equilibrium; a language exists as a ‘system of systems’, and maintains balance in order to function. At times, a catalyst will prompt a disequilibrium in the system, causing the language to adjust in order to reestablish equilibrium. There are several types of catalysts, ranging from “internal” weaknesses within the language due to psycholinguistic factors, such as difficulties coordinating verbal pronunciation with written word forms, to “external” influences from foreign languages or language learners. Both internal and external catalysts appear frequently throughout the history of the subjunctive from Latin into Romance. Chapters Two and Three of this thesis examine patterns of variability and stability in subjunctive forms and use in the evolution from Classical Latin into Modern French and Spanish. Chapter Two focuses on the formal and morphological evolution of the subjunctive verb system, beginning first with detailing the subjunctive inflections in Classical and Vulgar Latin, and then those in Old and Modern Spanish as well as in Old and Modern French. The direct 21 comparison of these forms leads to an analysis of patterns of change, identifying which modern verb forms are direct continuations of the original Latin and which have changed. The chapter concludes with hypothesizing regular patterns of change in the development of the subjunctive system, connecting observed shifts with existing theories of language evolution. Similarly, Chapter Three of this thesis centers on the pragmatic and semantic evolutions of the subjunctive mood, with an emphasis on the changes in psychosocial meaning over time. Beginning again with Classical and Vulgar Latin, the chapter works towards Modern Spanish and French through Old Spanish and French varieties by presenting snapshots of subjunctive use in different languages. Then, the patterns of variability and stability will be analyzed, with the material pragmatic changes linked to morphological shifts from the previous chapter. Chapter Four draws upon observations made in Chapters Two and Three as well as hypothesized patterns of change and stagnation in the subjunctive in order to predict the future development of the mood in the Romance family in general and in Spanish and French in particular. In addition to delineating and analyzing why languages change, this thesis aims to discuss theories of how languages change. This thesis relies on a diachronic approach to contrast the subjunctive in the languages at specific points in time, and while differences are apparent, it is difficult to trace back exactly how certain morphologies evolved. Many morphological developments in the subjunctive are attributed to simple changes in sound systems and pronunciation (as is generally the case in the Romance present subjunctive), while others are deeply intertwined with morphological shifts in the indicative (as is the case for the perfect subjunctive systems). Two interconnected models of language change are the “drag chain 22 reaction” and the “push chain reaction”; in the drag chain, a sound or piece of syntax “moves from its original place, and leaves a gap which an existing sound [or syntax] rushes to fill, whose place is in turn filled by another, and so on”, and in the push chain, the reverse occurs, where a sound or syntax “invades the territory of another, and the original owner moves away before the two sounds merge into one. The evicted sound in turn evicts another, and so on…” (Aitchison 2001, pg. 186). While these models are usually applied to sound shift paradigms, this thesis argues that the evolution of the subjunctive has operated in much of the same way, with morphological forms, especially within the perfect system, engaging in push and drag chains. Chain reactions and the restoration of morphological equilibrium are detailed in Chapter Two. French and Spanish, both prominent members of the Romance language family, obviously share a common ancestry and are similar in several morphological forms and pragmatic uses of the subjunctive mood. For example, the tense category of the present subjunctive in both Modern French and Modern Spanish has remained relatively unaltered and is directly descended from the present subjunctive found in Classical Latin. However, despite similarities, French and Spanish are often placed along a continuum of “conservatism”, or closeness to Latin, with Spanish categorized as more conservative in forms and uses of the subjunctive, whereas French is placed the furthest from Latin out of all five of the major Romance languages. Despite nominally preserving the original subjunctive categories from Classical Latin (present, perfect, imperfect, and pluperfect, see Table 1), the imperfect and pluperfect subjunctives are no longer used in Modern French, and several pragmatic functions of the subjunctive mood are increasingly expressed using the conditional. Some linguists theorize that the subjunctive will disappear entirely from the language, and even extrapolate that Spanish 23 is soon to follow. This thesis, on the other hand, contends that the subjunctive mood itself is neither losing meaning nor disappearing entirely, but rather increasing pragmatic significance with each decrease in morphological form. This theme is developed throughout Chapter Four, which discusses future evolutions of the Romance subjunctive system. One factor in predicting the next steps of the subjunctive mood is the examination of the relationship between nonnative language learners and the subjunctive, paying particular attention to the patterns of development in learners’ interlanguage, which are described in Chapter Five of this thesis. Language learners are a crucial component in linguistic evolution, and exercise a non negligible amount of influence over how language changes, a power which has only amplified in today’s digital age. As with many topics surrounding the subjunctive mood, researchers are divided on language learners’ subjunctive selection and its subsequent implications. This thesis acknowledges findings of lower subjunctive use, but maintains the position that the subjunctive has not diminished in pragmatic significance. Overall, this thesis attempts to organize and present a unified account of the development of the subjunctive mood from Classical Latin into Modern French and Modern Spanish, examining evolutions in morphological forms and pragmatic uses. It also attempts to identify specific processes of change as well as patterns of variability and stability in the subjunctive systems. This analysis, along with data reviewing interactions between language learners’ interlanguage and the subjunctive, allows for the prediction of further evolution of the Romance subjunctive systems. The central argument of the thesis supports the theory of the subjunctive as pragmatically significant and continually relevant, making up for reductions in morphological 24 forms with increases in psychosocial meaning. The conclusions found are then applied to the overarching questions of why and how languages change. 25 “‘They’ve a temper, some of them-- particularly verbs, they’re the proudest-- adjectives you can do anything with, but not verbs-- however, I can manage the whole lot of them!’” - Humpty Dumpty, from Through the Looking Glass by Lewis Carroll 2. THE MORPHOLOGICAL EVOLUTION OF THE SUBJUNCTIVE MOOD: FROM CLASSICAL LATIN TO MODERN ROMANCE This chapter endeavors to catalog and discuss the most salient of the changes to the subjunctive system: the morphophonemic evolution between Classical Latin and Modern Romance that propelled the development of the subjunctive verb endings themselves. At the morphological level, it is apparent that the subjunctive varies greatly between Latin, Spanish, and French. Tracing the histories of each mood system reveals not only patterns of variability, overlapping cycles of linguistic adjustment and readjustment, but also patterns of stability, linguistic elements that stand the test of time. An investigation of these language-specific patterns enriches present understandings of both the Romance language family and of language itself. Although languages are often seen as sets of rules, it will become clear that as a tool of communication, languages are often just as “alive” as their speakers. Languages grow and change, respond to the world around them as well as the multitudes they contain inside. Linguistic changes that have happened centuries ago continue to affect the development of language today. While at first glance it may seem that Latin is only the distant ancestor of Spanish and French, further study proves that the roots of language run deeper than originally thought. The following chapter will analyze variability and stability pattern types using the lens of established theories of linguistic change to make sense of the Classical past and pave the way for predicting the Modern future. 26 2.1 Latin Latin, the language of the Roman Republic and Empire, has been studied for centuries, both in its own right as a language but also as the direct linguistic predecessor of the Romance language family. As Penny (2002) states in his iconic book A History of the Spanish Language: Latin is the ancestor of Spanish (and, by definition, of all other Romance languages) in the sense that there is an unbroken chain of speakers, each learning his or her language from parents and contemporaries, stretching from the people of the Western Roman Empire two thousand years ago to the present population of the Spanish-speaking world. An alternative way of expressing the relationship between Latin and Spanish is to say that Spanish is Latin, as Latin continues to be spoken in parts of Europe, Africa and America. Similar claims are of course justified in the case of Portuguese, Catalan, French, Italian, Romanian, etc., and the main reason the term ‘Latin’ is not used for these various kinds of speech and writing is one of convenience: some forms of contemporary Latin (i.e. some Romance languages) have become mutually unintelligible and it is inconvenient to use a single label for mutually unintelligible forms of Language. (p. 4) Thus, the story of Spanish and French begins first with the story of Latin, which in turn finds its humble beginnings in an area of modern Italy previously known as Latium. History books and linguistic anthologies alike chart the rise of Latin from a local dialect to the language of prestige in an empire spanning across three continents and hundreds of years. 2.1.1 The Classical Latin Verb System By the first century C.E., what is now known as “Classical Latin” had emerged as the literary standard language of the educated elite in the late Roman Republic and into the early Roman Empire, its use stretching into the third century before a final evolution into “Late Latin”. Classical Latin was heavily prescribed; the linguistic “standard” was deliberately and meticulously maintained by writers and scholars of the era, who published works on “correct” grammatical usage and criticized the mistakes of others. Wealthy and elite circles were highly 27 educated, and linguistic norms were reinforced in the classroom. The lower classes, however, did not have access to education, and were relatively removed from explicit grammatical instruction. In Classical Latin, this juxtaposition between the speech of the aristocracy and that of the general population only grew with time, leading eventually to the emergence of “Vulgar Latin”. The grammar of Classical Latin had a rather rigid structure, and every sentence was a precise and logical construction. The systematic organization of Latin can be clearly observed in the verb; the verb system itself is a category of symmetrical subcategories, with each subcategory composed in turn of paradigmatic patterns. The verb system of Classical Latin can be described in terms of the following subcategories: Mood, Voice, Tense, Aspect, and Person. 2.1.1.1 Mood As discussed in the previous chapter, grammatical mood is the ability of a verb to express modality, or the speaker’s attitude toward an eventuality. Classical Latin makes use of three Mood categories: the indicative, used for indicating facts, certainties, and otherwise ‘real’ eventualities; the imperative, used for making commands; and the subjunctive mood, used for hypothetical or potential eventualities, wishes or impossible eventualities, as well as in certain subordinate clauses describing things like reported speech. 2.1.1.2 Voice Unlike the modern day Romance languages, Latin expresses grammatical voice synthetically. Broadly speaking, there are two types of verbal voice systems: the active system and the passive system. Active voice denotes that the verb is being done by the subject (e.g., “she buys a book”, in which the grammatical subject agreeing with the verb is “she”), whereas passive voice denotes that the verb is being done to the subject (e.g., “the book is bought by her”, in 28 which the grammatical subject agreeing with the verb is “the book”). Latin employs two infinitives for the purpose of expressing voice: the active infinitive, as in amāre “to love”; and the passive infinitive, as in amārī “to be loved”. A third class of verbs, called deponent verbs, are passive in form (i.e., have the appearance of the passive infinitive) but are active in meaning (i.e., signify the active voice). Both the active and passive infinitives can be conjugated into the full system of tenses. 2.1.1.3 Tense In addition to grammatical mood, grammatical tense was also presented in the first chapter. Grammatical tense refers to the time frame expressed by a given verb as compared to the frame of reference of the speaker. There are three time frames in Latin: the past, the present, and the future. 2.1.1.4 Aspect Grammatical tense can be further divided by verbal aspect. As well as locating a specific verb in a specific time frame (past, present, and future), Latin distinguishes whether or not the action expressed by the verb is complete or incomplete. For example, actions that happen at one specific point in time, have definite start and end points, or are not repeated all belong to the perfect aspect. On the other hand, actions that are habitual, repeated, or in progress all belong to the imperfect aspect. Tense and aspect combine to make six unique verb groups, also known as tenses. In Latin, the main six tenses are the present tense (“I walk”); the future tense (“I will walk”); the future perfect tense (“I will have walked”); the perfect tense, also known as the preterite (“I walked”/ “I have walked”), and the imperfect tense (“I was walking” / “I used to walk”). 29 Classical Latin establishes a morphological binary contrast between the “imperfect system” and the “perfect system” of verb tenses by using one verb ‘stem’ for the imperfect system, and another for the perfect. The ‘stem’ of the verb refers to the base part of the verb that carries the pragmatic meaning of the verb itself, without any person or tense markers. The imperfect stem is found by removing the infinitive conjugation ending (-āre, -ēre, -ere, or -īre). For example, the verb amāre “to love” has the imperfect stem of am-, while the verb dūcere “to lead” has the imperfect stem “dūc”. All of the imperfect tenses (the present, imperfect, and future) build off this imperfect stem; the tense and person endings are stuck directly behind the stem. On the other hand, the perfect stem of a verb is found from the first person singular preterite ending, also known as the third principal part (in English, this form would be the rough equivalent of “I did”), and removing the person marker from the conjugated verb form. In the case of amāre “to love”, which has the third principal part of amāvī “I loved”, the perfect stem is amāv-; dūcere “to lead” becomes dūxī “I led” which in turn yields the perfect stem dūx-. All of the perfect tenses (the preterite, pluperfect, and future perfect) begin with this perfect stem. 2.1.1.5 Person The final expression of a verb form is the person, or who is doing the action. In Latin, there are three persons: the first person, the second person, and the third person. These persons can be singular or plural, combining to a total of six possible persons: the first person singular (“I” in English); the first person plural (“we”); the second person singular (“you”) and plural (“you all”); and the third person singular (“he/she/it”) and plural (“they”). Every active tense in the Latin verbal system preserves the same set of person endings: the first person singular will always end in “m” or a vowel; the second person singular in “s”; the third person singular in “t”; 30 the first person plural in “mus”; the second person plural in “tis”; and the third person plural in “nt”. The following table presents these endings. Table 2: The Active Person Endings in Classical Latin Singular Plural First Person -m/vowel -mus Second Person -s -tis Third Person -t -nt 2.1.1.6 Verb Conjugation Classes (thematic vowels) Latin infinitives, the unconjugated forms of the verb that correspond to the English “to [verb]”, are divided into four main categories, also known as conjugations, based on thematic vowel sounds and stress. The first conjugation consists of verbs with active infinitives ending in -āre, as in the verb amāre “to love”. The second conjugation consists of verbs ending in -ēre, as in vidēre “to see”. The third conjugation is differentiated from the second by stress alone: these verbs end in -ere, with the stress pronounced on the syllable before the ending instead of the ‘e’, as in the verb dūcere “to lead”. A further subcategory of the third conjugation arises with the class of verbs with an additional theme vowel i, as in the verb capere “to capture”, which conjugates capiō as opposed to dūcō from dūcere. The fourth class, verbs ending in -īre, as in audīre “to hear”. The conjugation class of a verb determined the paradigmatic endings applied for each of the tenses. 2.1.1.7 The Classical Latin Subjunctive System Table 1: The Active Subjunctive Paradigm in Classical Latin 31 A S P E C T Incomplete Complete T E N S E Present Present ego amem - I love tū amēs - you love is/ea/id amet - he/she/it loves nōs amēmus - we love vōs amētis - you all love eī/eae/ea eament - they love Preterite ego amāverim - I [have] loved tū amāverīs - you [have] loved id/ea/id amāverit - he/she/it [has] loved nōs amāverīmus - we [have] loved vōs amāverītis - you all [have] loved eī/eae/ea amāverint - they [have] loved Past Imperfect ego amārem - I loved tū amārēs - you loved is/ea/id amāret - he/she/it loved nōs amārēmus - we loved vōs amārētis - you all loved eī/eae/ea amārent - they loved Pluperfect ego amāvissem - I had loved tū amāvissēs - you had loved is/ea/id amāvisset - he/she/it had loved nōs amāvissēmus - we had loved vōs amāvissētis - you all had loved eī/eae/ea amāvissent - they had loved Note: The regular inflections are bolded. Above is the full subjunctive paradigm for regular active first conjugation verbs in Classical Latin. Although the future and future perfect tenses exist in the indicative mood, they do not exist in the subjunctive, leaving only two time frames of reference: the present and the past. Taking verbal aspect into account, four tenses can be defined in the subjunctive mood: the present, the imperfect, the preterite, and the pluperfect, also known as the past perfect. 2.1.1.7.1 Present The present subjunctive paradigm in Latin is directly related to the present indicative, with one important difference: the “switch” in thematic vowels. The first conjugation -āre verbs can be distinguished in their subjunctive forms by the substitution of “e” for “a”. For example, 32 while in the indicative, the present of the first conjugation verb amāre is amat “he loves”, in the subjunctive it is amet. In the second, third, and fourth conjugations, the switch in thematic vowels is from “i” to “a”. In this case, the present third person singular form of the third conjugation verb dūcere switches from the indicative dūcit to the subjunctive dūcat “he leads”. This thematic vowel change pattern continues for the remaining person endings. 2.1.1.7.2 Imperfect The imperfect subjunctive in Classical Latin is relatively simple to form morphologically: the general person endings (“-m”, “-s”, “-t”, “-mus”, “-tis”, “-nt”) are added directly to the present active infinitive of the verb. Thus, the third person singular imperfect subjunctive form of the verb amāre is amāret, the first person plural imperfect subjunctive form of the verb dūcere is dūcerēmus, and so on. 2.1.1.7.3 Preterite As mentioned in section 2.1.1.4, the formation of the subjunctive preterite begins with the perfect stem, since the preterite belongs to the perfect system. After finding the perfect stem by removing the person ending from the third principal part (the first person singular indicative preterite), “eri” and the relevant general person ending are added. Therefore, the preterite subjunctive for the third person plural of amāre is amāverimus (amāv + eri + mus); the second person singular of dūcere is dūxeris (dūx + eri + s), and so on. 2.1.1.7.4 Pluperfect Much like the preterite, the pluperfect subjunctive forms also begin with the perfect verb stem. Following the stem are “sse” and the relevant general person ending. Thus, the pluperfect 33 subjunctive for the second person plural of amāre is amāvissetis (amāv + isse + tis); the first person singular of dūcere is dūxeris (dūx + isse + m), and so on. 2.1.2 The Vulgar Latin Subjunctive Although the term “Classical Latin” has remained a widely-accepted standard definition for a subsection of the Latin language corresponding to a set timeframe, over centuries of Latin scholarship the term “Vulgar Latin” has remained consistent only in terms of controversy. Earlier scholarship in the 19th century conceived of Vulgar Latin as the so-called ‘missing link’ between Classical Latin and the Romance language family, since discrepancies in vocabulary and phonetics between the two became objects of increasing study and interest in the linguistic community. The original idea of “Vulgar Latin” (from Cicero’s vulgaris sermo, “speech of the common people”, not a derogatory term) was that the less educated, poorer groups of people under Roman rule, being essentially excluded from the standards of Classical Latin, spoke their own form of Latin that differed so greatly from the usage of the literary elite that it indeed was an entirely new language, one which eventually spawned the entire Romance family. A second school of thought, one embraced by most modern scholars, acknowledges Vulgar Latin as “the set of all those innovations and trends that turned up in the usage, particularly but not exclusively spoken, of the Latin-speaking population who were little or not at all influenced by school education and by literary models” (Herman 2000:7). Although at first glance these definitions may seem identical, there are some important clarifications that must be taken into account. The first is that Vulgar Latin coexisted with Classical Latin, and resists chronological categorization; that is, Vulgar Latin came into existence with the written standard of the language (Classical Latin), and ceased to exist with the death of Latin as a spoken 34 language. The second is that, being defined as a spoken variety of Latin, Vulgar Latin “texts” do not exist per se. Writing itself necessitates certain conventions (letter systems, to name just one), and Vulgar Latin did not have expressly codified conventions. While evidence for Vulgar Latin has been collected through textual sources (graffiti, inscriptions, medieval writings etc), at best these sources can be described as “a text markedly influenced by Vulgar usage” (Herman 2000:8). Most importantly, unlike Classical Latin, Vulgar Latin was remarkably fluid and variable. Herman (2000) notes that: It needs to be understood, when undertaking the study of Vulgar Latin, that this is a set of highly complex and ever moving phenomena; it naturally changed over time, and the usage of the first century A.D. was considerably different from that of the sixth century and later; it also varied from place to place, and these geographical differences vary in importance at different times. In addition, Vulgar Latin undoubtedly had stylistic subvariants within itself, such as the jargons used in different technical spheres; thus it seems certain, for example, that the Vulgar usage of the Christian communities was not the same, particularly in vocabulary, but quite possibly in grammatical details as well, as the soldiers′ slang used in the same place at the same time. (pg. 8) Of course, the constant linguistic fluctuations and innovations of Vulgar Latin at any given time make the task of establishing a set grammar nearly impossible. The picture of the subjunctive mood presented here is partial at best, and is not assumed to represent the entirety of Vulgar Latin usage. Phonetic and morphological changes are perhaps easier to pinpoint, document, and analyze than changes in the subjunctive, since they appear with a greater frequency in the little historical evidence available. 2.1.2.1 Phonetic Change Two major categories of phonetic change occurred between Classical and Vulgar Latin: shifts in vowel sounds, and shifts in consonant sounds. Classical Latin had a symmetrical arrangement of 35 five basic vowels (a, e, i, o, and u). Each vowel could be pronounced as long or short, and each vowel sound was a distinct phoneme; that is, the difference in a long or short vowel distinguished different words, as is the case for malum with a short /a/ (“bad”) and malum with a long /a/ sound (“apple”). In total, there were ten total vowel sounds (two for each vowel). In some cases, these vowel sounds could change the tense of a verb. For this reason, correctly pronouncing the vowel sounds was incredibly important. However, in the first two centuries before the fall of the Roman Empire, there began a drastic change in vowel pronunciation (Herman 2000:28). The distinction between long and short vowels weakened until it collapsed entirely, leading to a period of orthographic confusion. In place of the system of vowel length, Vulgar Latin based sound distinctions on vowel quality (how open or closed the vowels are, wherein the more “closed” a vowel sound, the higher the relative position of the tongue during pronunciation). Herman (2000) further details that: The originally short vowel [/e/] had become an open /ε/, and the originally long vowel [ /e:/] had become a closed /e/ [...] In the case of the /a/, which is the most open vowel of all, relative quality did not become distinctive in this way; the short /a/ and the long /a:/ merged into a single phoneme /a/. The open version of the /i/ and the open version of the /u/, both originally short, did not survive as separate phonemes; in most places, originally short /i/ and originally long /e:/ merged into a single phoneme, the closed /e/, and in the same way, the back vowels originally short /u/ and originally long /o:/ merged into a single phoneme, the closed /o/; this was because in each case the mode of articulation of the two was similar, as was their sound. (ppg. 30-31) In the realm of consonant sounds, the changes relevant to the scope of this study are as follows: the loss of the word-final /m/ sound, the merger of the /b/ and /w/ sounds, and the increasing prevalence of syncope, or the deletion of an unstressed vowel in the middle of a word (Alkire & Rosen 2010: 28). Each of these changes in sound had an effect on the verb systems of Vulgar Latin, and on the subjunctive system in particular. 36 2.1.2.2 Morphological Changes Morphological changes, and adjustments to verb conjugations especially, followed some time after changes in vowel and consonant sounds. All of the information conveyed by a verb is encoded in the phonetic profile of the word itself; the verb stem (for example am- from amāre “to love”) conveys the inherent meaning of the verb (as in “love”), while the ending that is added onto the stem conveys additional qualifiers like mood, tense, aspect, person, and number. Thus, within the single word amat the interpreter (the one reading or listening) can identify an incredible amount of detail from two syllables. The changes included in the previous section led to further changes in a wide range of morphological areas, from verb conjugation classes to the verb tenses themselves. 2.1.2.2.1 Verb Conjugation Classes As previously described, Classical Latin categorizes verbs into four general verb conjugations based on thematic vowel sounds and stress; -āre, -ēre, -ere, and -īre. Due to years of subtle shifts in vowel pronunciation, some distinctions between the latter three classes blurred, and there arose general confusion as to which verbs belonged to which classes; it was not uncommon for the incorrect endings to be placed onto a verb after a misconception of conjugation class, and eventually some verbs switched conjugation classes entirely (Herman 2000: 70). Overall, however, the verb categories remained generally intact. 2.1.2.2.2 Aspect Classical Latin only distinguishes between imperfective and perfective aspects, and does not utilize a true perfect aspect. For example, as shown in Table 1, ego amāverim is given two alternate translations in English: “I loved” and “I have loved”. In Vulgar Latin, a construction 37 was appropriated using the verb habere “to have” and the past participle of the verb (for example, amātus “loved” from the verb amāre “to love”) to distinguish this contrast. 2.1.2.2.3 Tenses Table 3: The Active Subjunctive Paradigm in Vulgar Latin A S P E C T Incomplete Complete T E N S E Present Present ego áme - I love tū ámes - you love is/ea/id amet - he/she/it loves nōs amémus - we love vōs amétis - you all love eī/eae/ea áment - they love Perfect ego habeam amātu - I have loved tū habeās amātu - you have loved is/ea/id habeat amātu - he/she/it has loved nōs habeāmus amātu - we have loved vōs habeātis amātu - you all have loved eī/eae/ea habeant amātu - they have loved OR ego amā(ver)im - I [have] loved tū amā(ve)rīs - you [have] loved id/ea/id amā(ve)rit - he/she/it [has] loved nōs amā(ve)rīmus - we [have] loved vōs amā(ve)rītis - you all [have] loved eī/eae/ea amā(ve)rint - they [have] loved Past Imperfect ego amā(vi)ssem - I loved tū amā(vi)sses - you loved is/ea/id amā(vi)sset - he/she/it loved nōs amā(vi)ssēmus - we loved vōs amā(vi)ssētis - you all loved eī/eae/ea amā(vi)ssent - they loved Pluperfect ego habērem amātu - I have loved tū habērēs amātu - you have loved is/ea/id habēret amātu - he/she/it has loved nōs habērēmus amātu - we have loved vōs habērētis amātu - you all have loved eī/eae/ea habērent amātu - they have loved OR ego amā(vi)ssem - I had loved tū amā(vi)ssēs - you had loved is/ea/id amā(vi)sset - he/she/it had loved nōs amā(vi)ssēmus - we had loved vōs amā(vi)ssētis - you all had loved 38 eī/eae/ea amā(vi)ssent - they had loved Note: The regular inflections are bolded. The above chart illustrates an estimation of the active subjunctive paradigm in Vulgar Latin. Of course, as Vulgar Latin has no widespread standardization or unification across time and place, the actual endings varied greatly. The paradigm displayed in this study represents the greatest level of contrast between Classical and Vulgar Latin; in reality, the process of linguistic differentiation took centuries. 2.1.2.2.3.1 Present The present tense of active subjunctive in Vulgar Latin is relatively unaltered compared to the Classical Latin form. The conjugation process remains the same; the only discrepancies arise in vowel sounds and stress (shifting from long to short), and the loss of the final m in the first person singular. Both of these changes are reflective of the phonetic evolution discussed earlier in the chapter. 2.1.2.2.3.2 Imperfect The morphology of the imperfect subjunctive in Vulgar Latin diverges significantly from that of Classical Latin. The original morphology fell out of use due to difficulties in pronunciation (Herman 2000:79), and was replaced by two sets of competing endings recycled from other preexisting ones. The first set, beginning with amāveram, was originally the morphological paradigm for the pluperfect indicative, while the second, beginning with amāvissem, was originally the pluperfect subjunctive. Due to the prevalence of syncope, both sets of forms could be pronounced and written without the middle v + vowel (for example, amāveram is shortened to amāram, amāvissem to amāssem and so on). 39 2.1.2.2.3.3 The Perfect System The birth of the periphrastic (two-word) perfect system in Vulgar Latin was a relatively gradual process of linguistic evolution, originating from a popular misinterpretation of the habere + past participle construction. The system starts as an alternative paradigm for the perfect indicative, wherein habere is conjugated in the indicative and combined with the past participle. The periphrastic (two-word) and the synthetic (one-word) indicative perfects coexisted as equivalent terms, until eventually the periphrastic construction took on a more nuanced meaning with a true perfect aspect (have done something as opposed to did something). Once the new present perfect had taken hold in the indicative system, it spread to the subjunctive. The so-called perfect system in Vulgar Latin has two tenses: the preterite and the pluperfect. The preterite (eventually the present perfect) is formed by conjugating habere into the present tense and adding the past participle, while in the same way the pluperfect (or past perfect) is formed by conjugating habere into the perfect tense and adding the past participle. This conjugation process is analogous to that of the English perfect system. 2.2 Spanish Subjunctive Morphological Evolution The Latin language arrived in the Iberian Peninsula around 218 B.C.E. with the beginning of the Second Punic War. Roman troops eventually captured the Carthaginian capital of Cádiz, and thus began the slow but steady process of colonization and latinization. However, notes Ralph Penny (2002) in A History of the Spanish Language: The pace of latinization is probably correlated with geographical distance from the ‘educated standard’ of the ‘average’ Latin spoken at any given date. The factors which encouraged rapid latinization (close contact with central Italy, urbanization, good road communications, the consequent fostering of trade, etc.) are the same factors which encouraged the use of forms of Latin which were closer to the prestigious end of the sociolinguistic spectrum... It is therefore likely that the ‘average’ Latin spoken by people in the remoter, less developed, parts of the Peninsula was considerably further from the 40 prestige norm (that of upper-class Rome) than was the speech of the eastern and southern cities. This factor is particularly relevant to the history of Spanish, since Spanish has its geographical roots in what is now the northern pan of the province of Burgos, an area of the northern meseta which was remote from the centres of economic activity and cultural prestige in Roman Spain, which was latinized fairly late, and where the Latin spoken must consequently have been particularly remote from the prestige norm (that is, particularly ‘incorrect’) at the time of the Roman collapse. With the end of the Roman state came the effective removal of the linguistic model towards which, however distantly and ineffectually, speakers strove to adhere, so that any ‘incorrect’ features of local speakers were likely to be perpetuated (unless challenged by some other prestige model, which was not to be the case in the Burgos area). (pp. 9-10) Of course, these same factors encouraging linguistic variation between Rome and her territories also encouraged a wide variety of language differences within the territory of the Iberian Peninsula itself. At this point scholars begin to speak of Hispano-Romance, an emerging collection of dialects innovating upon the standard Latin of colonization. In the absence of the Roman Empire’s colonial shadow, the Iberian Peninsula was broken into smaller territories and city states, some Christian, others Muslim. A period of political soul searching began; the kingdom of Castille emerged from the north as a major force unifying other ‘Latin’ kingdoms against the Islamic rule to the south. Thus the Romance spoken in Castille became the Hispano-Romance variety of linguistic prestige, and from it originated the standards which shaped what is today known as the Spanish language. 2.2.1 The Old Spanish Subjunctive The Old Spanish, referred to also as Medieval Spanish or Medieval Hispano-Romance, surfaces beginning in the tenth century and remains until approximately the end of the fifteenth century (Dworkin 2018: 5). 41 2.2.1.1 Phonetic Change Old Spanish has a five vowel system (/a/, /i/, /e/, /o/, /u/), a reduction from the seven vowel sounds in Vulgar Latin, although written evidence indicates that pronunciation waivered: orthographically, /e/ and /i/ as well as /o/ and /u/ were often confused. 2.2.1.2 Morphological Changes Old Spanish continued the evolutionary process that began in Latin; further adjustments to the phonetic system of the language prompted changes in the morphological expression of the subjunctive system. 2.2.1.2.1 Verb Conjugation Classes The shift in vowel sounds impacted the categorization of verb infinitives, which was based on phonetic distinction. As previously mentioned, Classical and Vulgar Latin divide verbs into four conjugation classes based on vowel sound, with the first conjugation reserved for verbs ending in -āre, the second for verbs ending in -ēre, the third for -ere, and the fourth for those verbs ending in -īre. Mergers in vowel sounds beginning in Classical Latin ultimately culminated in the merging of the second and third conjugation classes, so that Old Spanish verbs used a three class system of -ar, -er, and -ir verbs. Depending on factors such as word stress, verbs moved from the third conjugation to the second or from the third to the fourth, although in rare cases verbs from other classes have also switched. 2.2.1.2.2 Aspect While both Classical and Vulgar Latin base the verbal system on a binary distinction between perfective and imperfective aspect, Old Spanish introduces a third aspect to the mix: the perfect. In simple terms, the perfect aspect system relies on the auxiliary verb aver (from Latin habere, 42 later haber in Modern Spanish) and the main verb as a past participle, a rough equivalent of the English system with the verb “to have done”. Although the grammatical structure dates back to Classical Latin, in Old Spanish it is realized as a complete verbal system. 2.2.1.2.3 Tenses Table 4: The Active Subjunctive Paradigm in Old Spanish A S P E C T Incomplete Complete T E N S E Present Present ame - I love ames - you love ame - he/she/it loves amemos - we love amedes - you all love amen - they love Perfect haya amado - I have loved hayas amado - you have loved haya amado - he/she/it has loved hayamos amado - we have loved hayades amado - you all have loved hayan amado - they have loved Past Imperfect amasse / amara - I loved amasses / amaras - you loved amasse / amara - he/she/it loved amássemos / amáramos - we loved amássedes / amárades - you all loved amassen / amaran - they loved Pluperfect oviesse / oviera amado - I had loved oviesses / ovieras amado - you had loved oviesse / oviera amado - he/she/it had loved oviessemos / ovieramos amado - we had loved oviessedes / ovierades amado - you all had loved oviessen / ovieran amado - they had loved Future Future amare - I will love amares - you will love amare - he/she/it will love amáremos - we will love amáredes - you all will love amaren - they will love Note: The regular inflections are bolded. 43 2.2.1.2.3.1 Present The present subjunctive remains basically congruent to that of Classical and Vulgar Latin. 2.2.1.2.3.2 Imperfect Here, Old Spanish follows the lead of innovations in Vulgar Latin; the imperfect subjunctive forms were replaced by the pluperfect subjunctive, and the newly imperfect forms were subsequently syncopated (shortened). Old Spanish derives its imperfect subjunctive set directly from those of Vulgar Latin. Towards the end of the Old Spanish period, a second set of forms, those from the imperfect indicative, were also used as imperfect subjunctives. 2.2.1.2.3.3 Future A novel innovation in Old Spanish, and indeed in Romance as a whole, is the future subjunctive. These forms were recycled using the morphology from the previously defunct Latin future perfect indicative and the perfect subjunctive (Alkire & Rosen 2010: 132). 2.2.1.2.3.4 The Perfect System As previously mentioned, the perfect subjunctive system descending from Vulgar Latin is a collection of two periphrastic tenses, using a morphological formula of conjugating the verb aver (from Latin habere) into either the subjunctive present or imperfect and adding the past participle of the main verb to create either the present perfect or pluperfect subjunctive, respectively. Old Spanish also used the verb ser (to be) as an auxiliary to form these tenses; aver was used for constructions with a transitive verb (having a direct object), while ser was used for constructions with an intransitive verb (having no direct object). Both constructions appear relatively infrequently in Old Spanish source material. 44 2.2.2 The Modern Spanish Verb System Scholars pinpoint the beginning of the Modern Spanish period almost six hundred years ago, with the publication of Antonio de Nebrija’s iconic work Gramática de la lengua castellana 1492, the first grammar in Spanish. First utilized as a political tool to unify Spain under the Catholic kingdom of Castile and León, the Spanish language traveled overseas and was used in the colonization of Latin America. Although Modern Spanish has obviously undergone linguistic evolution since that time, most features remain consistent. 2.2.2.1 Phonetic Change In terms of consonant shifts from Old to Modern Spanish that are relevant to the scope of this study, the most prominent is the merging of /b/ and /ß/. While in Old Spanish the distinction between “b” and “v” was consistent, in Modern Spanish these sounds have merged completely. Other primarily orthographic changes include the reduction of “ss” to “s”, especially noticeable in the imperfect subjunctive paradigm, as well as the deletion of the “d” in the second person plural forms across the entire Spanish verbal paradigm due to vowel stress shift. An additional noticeable change appears with the verb aver, used in the formation of the perfect system, which was relatinized after the original etymology (the Latin verb habēre). 2.2.2.2 Morphological Changes Although the use of the pluperfect indicative endings (amara) as a new paradigm for the imperfect subjunctive was introduced in Vulgar Latin and survived into Old Spanish, the most used set of endings were those formerly from the pluperfect subjunctive (amasse/amase). The two paradigms existed in competition with one another, and Modern Spanish saw the eventual rise and domination of the amara forms. Both sets are still used today, but the amasse forms are 45 much less common, and perceived as archaic among modern speakers. The future subjunctive, while only emerging in Old Spanish, nears extinction in the Spanish of today. It, as well as its perfect counterpart, are seldom used except in legal or formal texts. In the perfect system, Modern Spanish completes the imperfect/perfective binary with the introduction of the future perfect subjunctive. The use of haber reaches majority status, and is used with both transitive and intransitive constructions; forms of ser are no longer found in perfect constructions. 2.2.2.2.2 Tenses Table 5: The Active Subjunctive Paradigm in Modern Spanish A S P E C T Incomplete Complete T E N S E Present Present ame - I love ames - you love ame - he/she/it loves amemos - we love améis - you all love amen - they love Perfect haya amado - I have loved hayas amado - you have loved haya amado - he/she/it has loved hayamos amado - we have loved hayáis amado - you all have loved hayan amado - they have loved Past Imperfect amasse / amara - I loved amasses / amaras - you loved amasse / amara - he/she/it loved amásemos / amáramos - we loved amaseis / amarais - you all loved amasen / amaran - they loved Pluperfect hubiese / hubiera amado - I had loved hubieses / hubieras amado - you had loved hubiese / hubiera amado - he/she/it had loved hubiésemos / hubiéramos amado - we had loved hubieseis / hubierais amado - you all had loved hubiesen / hubieran amado - they had loved Future Future amare - I will love Future Perfect hubiere amado - I will have loved 46 amares - you will love amare - he/she/it will love amáremos - we will love amareis - you all will love amaren - they will love hubieres amado - you will have loved hubiere amado - he/she/it will have loved hubiéremos amado - we will have loved hubiereis amado - you all will have loved hubieren amado - they will have loved Note: The regular inflections are bolded. 2.2.2.2.2.1 Present The present subjunctive remains fairly stable from Old Spanish to Modern Spanish, and indeed from Classical Latin to present day, with the exception of the second person plural (vosotros), wherein stress shift in vowels has caused the gradual omission of the intervocalic consonant (“t” or “d”). 2.2.2.2.2.2 Imperfect As previously mentioned, the two sets of inherited forms for the imperfect subjunctive have changed little, with the exception of minor orthographic adjustments. Pragmatically, however, whereas in Vulgar Latin and Old Spanish the paradigm recycled from the pluperfect subjunctive was most common, in Modern Spanish the forms originating in the pluperfect indicative are preferred. 2.2.2.2.2.3 Future The future subjunctive, an innovation dating back only to Old Spanish, passes relatively unchanged into Modern Spanish. Modern Spanish, in turn, innovates the future perfect subjunctive following the haber + past participle template used in constructing the other perfect tenses. 47 2.2.2.2.2.4 The Perfect System The most prominent evolution in the perfect system, besides the invention of the future perfect subjunctive, is the remodeling of oviesse / oviera to hubiese / hubiera. The “h” is reintroduced in the relatinization process of the verb haber; the “v” is replaced by a “b” for the same reasons and mostly likely aided by the phonetic merger of /b/ and /ß/. Modern Spanish removes most double consonants from the standard orthography, causing “ss” to become a singular “s”. The switch from “o” to “u” is probably related to the general phonetic confusion in the Old Spanish vowel merger. 2.3 French Subjunctive Morphological Evolution The story of French begins much like the story of Spanish. Although Julius Caesar's infamous conquest of Gaul began in 58 B.C.E., the Latin colonization of the territory known today as France began as early as 121 B.C.E., when Roman settlers began arriving in the area of Province. The settlers were not inhabiting empty land, however: the Gauls, a collection of Celtic tribes, were already there. The beginnings of the differentiation process between the “standard” Latin and the eventual romance dialects proceeding French can be traced back to the contact between the Romans and the Gauls. As was the case with Hispano-Romance and Spanish, the fall of the Roman empire isolated the varieties of Latin, leading to further development of the linguistic systems in the now separate territories. An additional complicating circumstance in the evolution of Gallo-Romance was the invasion of the Franks, who took control of Gaul. The Frankish rule did not, however, extend to complete linguistic dominion; while the Frankish language left deep footprints, it never replaced the ever-evolving Gallo-Romance. The emergence of Old French, a 48 collection of dialects rather than a new linguistic standard, is tied closely with the ascension and legacy of Charlemagne and the political prowess of the Christian church. The Middle Ages saw Old French flourish and slowly fade to Middle French in the twelfth century. Hundreds of years later in the sixteenth century Modern French took shape. 2.3.1 The Old French Subjunctive The subjunctive system in Old French was much more productive than that of Modern French, in terms of both morphological diversity and overall implementation of the mood. 2.3.1.1 Phonetic Change French, long considered the Romance language that has evolved the furthest from the common Latin ancestor, has an extensive history of phonetic change in vowels and consonants. Of these innovations, some are particularly relevant to the evolution of the subjunctive mood. For example, when stress on the verb stem (which appears in the first person singular, the second person singular, and the third person singular in plural) coincided with a “free” or open vowel, the diphthongization of that vowel would subsequently occur (Rickard 1989: 52): although the infinitive paradigm presented is for the verb amer “to love”, several forms begin with the diphthong “ai” instead of the expected “a” (visible in Table 5). Also noticeable in the table are the several variations in spelling; the merging of vowel sounds lead to general orthographic confusion, as the same sound could be written several ways. A third relevant development is the gradual yet dramatic reduction of the Vulgar Latin second person plural “ētis” to “ez”. 49 2.3.1.2 Morphological Changes The connection between Vulgar Latin and Old French, while appearing tenuous on the surface, is in reality quite direct on the morphological level. Accounting for several major changes which are detailed below, the two subjunctive systems are similar indeed. 2.3.1.2.1 Verb Conjugation Classes Similarly to Spanish, changes in vowel systems and word stress caused a consolidation of the verb conjugation classes. From the original four of Classical and Vulgar Latin, Old French creates three: the first, reserved for verbs ending in -er; the second, for verbs ending in -ir; and the third for verbs ending in -re. 2.3.1.2.2 Aspect Old French retains the binary distinction between perfective and imperfective aspects upon which the Latin verb system was based. Old French, like Classical and Vulgar Latin, uses two verb stems to form tenses; the present stem to form the imperfective tenses, and the perfect stem to form the perfective ones. 2.3.1.2.3 Tenses The following table presents the active subjunctive paradigm for regular verbs belonging to the first conjugation class using the verb amer “to love”. Nominally, all tenses categories are preserved from Latin, as well as the present-past and incomplete-complete (imperfective-perfect) matrices. The morphological forms themselves, however, have developed. Table 6: The Active Subjunctive Paradigm in Old French A S P E C T Incomplete Complete Present Present Perfect 50 T E N S E j’aim - I love t’ains - you love il aint - he/she/it loves nos amons - we love vos amez - you all love il aiment - they love j’aie amé - I have loved t’aies amé - you have loved il aie/aiet/ait amé- he/she/it has loved nos aiiens/aions amé - we have loved vos aiiez amé - you all have loved il aient amé - they have loved Past Imperfect j’amasse - I loved t’amasses - you loved il amast - he/she/it loved nos amissons/amissiens - we loved vos amissoiz/amissez/amissiez - you all loved il amassent - they loved Pluperfect j’eüsse/oüsse amé - I have loved t’eüsses/oüsses amé- you have loved il eüst/oüst amé- he/she/it has loved nos eüssons / eüssiens / oüssons / oüssiens amé - we have loved vos ëussoiz / eüssez / eüssiez / oüssoiz / oüssez / oüssiez amé - you all have loved il eüssent / oüssent amé - they have loved Note: The regular inflections are bolded. 2.3.1.2.3.1 Present Accounting for phonological change, the present subjunctive in Old French corresponds directly to that of Vulgar Latin. The present subjunctive is formed using the present stem, which is found by conjugating the verb into its first person plural indicative form and removing the person ending (for example, the verb amer becomes amons, for which the present stem is “am-”). The person endings are added to this stem (nothing, s, t, ons, ez, ent). As previously mentioned, the stress pattern in the endings (where the root is stressed, every form but first and second person plural) causes the vowel “a” to morph into the diphthong “ai”. A second spelling change occurs in the second and third person singular forms; where an “m” is expected and “n” is written. This is due mostly to difficulty pronouncing the consonant clusters “ms”/“mt”. 51 2.3.1.2.3.2 Imperfect Unlike with Spanish, the imperfect subjunctive in Old French has only one paradigm, which descends from the Vulgar Latin (syncopated) pluperfect subjunctive set. Some forms, notably the first person plural and the second person singular and plural, went through periods of further phonetic contraction, resulting in the forms illustrated in the Table. This syncopation was not particular to the imperfect subjunctive, but indeed all endings for first and second person plural changed from the Latin “mus” to the competing Old French forms “ons” or “iens” and from “tis” to “ez”, “oiz”, or “iez”. 2.3.1.2.3.4 The Perfect System The perfect system, composed of the perfect and pluperfect tenses, are both formed using the auxiliary verb avoir, the Old French evolution of the Latin verb habere “to have”, with the past participle. This perfect system is equivalent grammatically to those implemented in late Vulgar Latin and in Old and Modern Spanish. The perfect uses the present tense of avoir (hence why the perfect tense is also known as the “present perfect tense”) with the past participle, which in the case of the verb amer “to love” is amé “loved”. The pluperfect, also known as the past perfect tense, is formed with the imperfect tense of avoir. The imperfect paradigm of avoir was greatly affected by the orthographic confusion in the Old French era; several spellings of the same verb form coexisted and competed for standard legitimacy. 2.3.2 The Modern French Verb System If Old French could be characterized as a period of linguistic free reign with incredible diversity and experimentation in language and dialects, Modern French could be construed as the exact opposite. In the early seventeenth century, the era of linguistic standardization had begun. 52 Writers, grammarians, scholars, and politicians decided that explicit rules for the language were in order; the dialect spoken around the political and economic center of Paris was chosen as the new standard. Grammars and linguistic critiques were published, schooling eventually became mandatory, and metalinguistic awareness grew to an all-time high. The formation of the Académie française for the standardization of French in 1635 marked a turning point in the linguistic rhetoric of the state. The ascension of France as a global power drove French to become the global language of prestige, and any “blemishes” on the language were systematically erased. In a relatively short period of time, the language became policed from the top down, and was shaped into a language that many believed was superior to all others. In fact, notes A History of the French Language, “By the time of the Revolution, the grammar of French had been codified to the last detail, and it was a codification which was to prevail, with very few modifications, down to the present day” (Rickard 1989: 100). 2.3.2.1 Phonetic Change Much may be said on the topic of the phonetic change of the French language, but most points lie outside of the scope of this study. In brief, spellings were standardized both to reflect the existing pronunciations and to reconnect etymologically with Latin. In Table 6 below, visible adjustments include the paradigm leveling switch from “a” to “ai” in all forms, as well as the orthographic change from “ii” to “y”. Several forms of the third person singular delete the residual “s” in front of the person marker “t”; in its place, the preceding vowel receives a circumflex accent. In the first and second person plural forms, an “i” is introduced before the regular endings, and “ons” becomes “ions” as “ez” becomes “iez”. 53 2.3.2.2 Morphological Changes Morphologically, the path from Old to Modern French is straightforward. Certain verb forms were remodeled to fit more consistently in the more common conjugation patterns. A conspicuous change in the subjunctive morphology is the renovation of the present subjunctive of the verb dire “to say”, wherein an “s” is introduced (for example, die becomes dise). 2.3.2.2.1 Verb Conjugation Classes Modern French retains three conjugation classes. As in Old French, the “-er” class derives from the first conjugation of Latin verbs ending in “-āre”. The second class consists of verbs ending in “-ir” or “-ïr”. The third class, often described as the class for verbs ending in “-re”, also includes a few “-ir” verbs, as well as irregular verbs like aller “to go” and envoyer “to send”. The first class contains the greatest quantity of verbs by far. 2.3.2.2.2 Aspect The aspectual system based on the distinction between perfective and imperfective aspects remains unchanged from Classical Latin. 2.3.2.2.3 Tenses Table 7: The Active Subjunctive Paradigm in Modern French A S P E C T Incomplete Complete T E N Present Present j’aime - I love t’aimes - you love il aime - he/she/it loves nous aimions - we love vous aimiez - you all love ils aiment - they love Perfect j’aie aimé- I have loved t’aies aimé - you have loved il ait aimé- he/she/it has loved nous ayons aimé - we have loved vous ayez aimé - you all have loved ils aient aimé - they have loved 54 S E Past Imperfect j’aimasse - I loved t’aimasses - you loved il aimât - he/she/it loved nous aimassions - we loved vous aimassiez - you all loved ils aimassent - they loved Pluperfect j’eusse aimé - I have loved t’eusses aimé- you have loved il eût aimé- he/she/it has loved nous eussions aimé - we have loved vous eussiez aimé - you all have loved ils eussent aimé - they have loved Note: The regular inflections are bolded. 2.3.2.2.3.1 Present The subjunctive present in Modern French is formed almost identically to that of Old French. First, the present stem is found by conjugating the verb into the third person plural present indicative and dropping the “ent” person ending. Then the relevant subjunctive ending is added to the stem. 2.3.2.2.3.2 Imperfect The imperfect subjunctive, today used only in literary texts, has a more complicated conjugation process. The tense has been linked with the passé simple tense (a preterite) since Vulgar Latin; both the imperfect subjunctive and the preterite are formed with the perfect stem (note that the imperfect subjunctive in Vulgar Latin recycles morphology from the pluperfect subjunctive, a perfect tense). Therefore, any changes to the French passé simple are duly reflected in the imperfect subjunctive (Alkire & Rosen 2010: 161). In Modern French, the conjugation of the imperfect subjunctive further depends on the class of the verb. For first conjugation verbs ending in “-er”, the third person singular passé simple form is used as a stem to which the person endings of the imperfect subjunctive are added. For second and third conjugation verbs, the final 55 “t” in the third person singular passé simple form must be dropped before adding the imperfect subjunctive endings. 2.3.2.2.3.3 The Perfect System As in late Vulgar Latin and Old French, the perfect subjunctive system is periphrastic (two-words) and composed of two tenses: the preterite and the pluperfect. The preterite, also known as the past subjunctive, is formed by conjugating an auxiliary verb into the present subjunctive using the relevant person ending and adding the past participle of the main verb. If the verb is transitive and takes a direct object, the auxiliary verb used will be avoir “to have”. If the verb is intransitive or pronominal (reflexive), the auxiliary verb used will be être “to be”. Similarly, in the pluperfect subjunctive, the auxiliary verb is conjugated into the imperfect subjunctive and the past participle is then added. 2.4 Patterns of Variability and Stability Now that the linguistic changes in the subjunctive paradigms have been thoroughly documented, it is possible to analyze the specific changes themselves as part of a larger, overarching pattern of language evolution. We begin with the analysis of patterns of stability, identifying what aspects of the subjunctive mood remained constant over the centuries and theorizing possible explanations for this cross-linguistic continuity. Next, the same approach is taken for the patterns of variability. Finally, the patterns and theories described in the previous sections are discussed from the perspective of existing theories of linguistic change. 56 2.4.1 Patterns of Stability At first, it may seem counterintuitive to devote effort to study linguistic elements that have not changed, but the investigation of patterns of stability can reveal valuable information on what triggers language evolution. The evolutionary path of the subjunctive mood from Classical Latin into modern Romance has involved few detours; several elements of the language have been consistent for centuries, both intra- and inter-linguistically. For example, in each of the three languages studied, the available tenses in the subjunctive remained the same. Despite changes in tense morphology, the four major tense categories (present, imperfect, preterite, pluperfect) were constant, along with the vast majority of person/number endings in the verb paradigms. The present subjunctive paradigm in particular persevered remarkably unchanged-- the conjugation process in all three languages is incredibly similar, beginning with the present stem and adding the ‘opposite’ endings from the present indicative, with the exception of French -er and -re verbs, which use the ‘same’ endings-- and the imperfect paradigm also demonstrated high levels of consistency. Vulgar Latin, French, and to some extent Spanish all retained the aspectual organization of the verbal system-- they kept the distinction between imperfective and perfective aspect, and continued the use of imperfective and perfective stems. The differentiation between these stems ‘links’ certain tenses together across the mood systems; all of the imperfective tenses are united, as are the perfectives. The perfective and imperfective symmetry originating in Latin, wherein all imperfective tenses (in the subjunctive, the present and the imperfect, and in the indicative, the present, imperfect, future) are conjugated with the imperfective stem from the infinitive, and all perfective tenses (the perfect and pluperfect in the subjunctive, and the indicative perfect, pluperfect, and future perfect) are conjugated with the perfective stem from 57 the third principle part (first person singular perfect active indicative). Note that in Classical Latin, the present and imperfect subjunctive forms can be broken down into ama + ending, while the perfect and pluperfect being with amav + ending. Although syncopation shortened the distinctive perfective marker “v” in the perfect and pluperfect subjunctives, the bond between the perfective tenses remained strong. When the pluperfect subjunctive paradigm moved to the imperfect tense, morphologically, the verb forms remained “perfective”, which explains why changes in the French indicative passé simple (the synthetic perfect) are reflected in the imperfect subjunctive; the indicative perfect determines the perfective stems, and when the perfective stems change, so do all the perfective forms. It can also be argued that the verb conjugation classes showed signs of stability, at least in terms of dividing verbs into separate classes determining morphological treatment. Furthermore, the verb conjugations all continued to be determined off of themed vowels (ā, ē, e, and ī in Latin; a, e, and i in Spanish; and e, i, and ə in French). Phonologically, some verb paradigms have preserved the original stress systems from Classical Latin, preserving morphemic patterns in verb paradigms. Although major shifts in pronunciation occurring between Classical Latin and Modern Romance altered most consonants and all but a few vowels, tendencies of conservation appear to be stronger within the individual languages; after a period of transformation where one language starts to develop into another, phonology and morphology seem to ‘settle in’, whether naturally or through conscious language planning. For example, diphthongs emerging out of Proto-Romance are retained-- note the diphthong “ie” in Old Spanish oviesse continue into the Modern Spanish hubiese, and the diphthong “ai” in Old French aim continue into Modern French aime. Changes still occur, but 58 each of these ‘non-developments’ reveals an important detail in the larger picture of linguistic evolution. 2.4.2 Patterns of Variability Most noticeable in this study have been the components of the subjunctive mood that have changed from one language to another or even within a language itself. On a phonological level, a crucial shift was the vowel merger that began in early Vulgar Latin, wherein the original ten vowel sounds in Classical Latin were gradually reduced to seven sounds. In Spanish, there was an additional reduction to five primary vowel sounds; In French, the seven sounds were expanded to around 4 nasal vowels and 12 oral ones due to processes like nasalization, palatalization, and lenition of consonants. Vulgar Latin, Spanish, and French also underwent separate processes of diphthongization, experimenting with combinations of vowel sounds. Some sounds that were originally pronounced without a diphthong gradually acquired one, especially in the case of French. In the realm of consonants, changes that began in Vulgar Latin greatly affected the development of the entire Romance subjunctive system. Shifting pronunciations of the /r/ sound, as well as the phonetic merger of /b/ and /w/, complicated the Classical Latin subjunctive paradigms. As previously discussed, several tense morphologies fell out of use entirely, especially those that became too similar in pronunciation. French and Spanish had continuous phonological evolution, but those shifts that did occur had minor effects on the subjunctive system. Two significant developments in the morphological history of the subjunctive are the shifts in stress patterns and the expansion of the periphrastic perfect system. In both Classical 59 and Vulgar Latin, the practice of syncope (the deletion of unstressed syllables in the middle of a verb) became prominent in speech and writing. Commonly syncopated were the verb forms with the perfect marker “vi” or “ve”. For example, amāveram is shortened to amāram, amāvissem to amāssem and so on. The morphological paradigms of the Romance subjunctive, and the imperfect subjunctive in particular, evolved from these syncopated Latin forms. Word stress also led to changes in spelling or even the shortening of morphological endings, as is the case in French, where the Classical Latin amētis /aˈmetɪs/ gradually became the Modern French aimiez /ɛˈmje/. The two-word perfect system is one of the most studied and debated topics in Romance morphology. The structure of habēre “to have” or esse “to be” with the past participle did exist in early Classical Latin, but was esoteric and only used in the present indicative form. When the simple one-word perfective forms were weakened by phonetic variations, the periphrastic perfect grew in popularity. Its use soon spread from the present indicative to the present subjunctive. Gradually, the entire periphrastic perfect paradigm was completed in both the indicative and the subjunctive, and the simple forms were mostly replaced or recategorized into different tenses, as in the case of the pluperfect subjunctive morphology moving to the imperfect subjunctive tense. French and Spanish each ‘inherited’ the periphrastic ‘two-word’ perfect system as well as the synthetic ‘one word’ system. In Spanish, both systems remain in use today, but within the two-word periphrastic perfect system, only haber, the equivalent of habēre “to have”, is employed as an auxiliary verb. In some varieties of Modern Spanish, the use of the periphrastic present perfect is restricted, and in others it is not used at all. In French, on the other hand, the 60 simple perfects are rarely used (deemed a ‘literary tense’), and the periphrastic perfect continues to operate on a transitive-intransitive verb binary between avoir “to have” and être “to be”. 2.4.3 Theories of Linguistic Change It would be easy to stand at the far edge of history and retrospectively ascribe certain narratives to the histories of language. Popular belief, clung to by so-called linguistic and cultural ‘purists’, bemoans the evolution of language as a process of simplification, a great ‘dumbing-down’ of linguistic structures. In this perspective, any change in language is seen as negative, and that ‘modern’ linguistic developments represent the rapid disintegration of the sanctity of language. The truth, ironically, is much more complex than a downward spiral. Several factors contribute to morphological change in a language. Aitchison (2001) identifies two major categories of linguistic development: “on the one hand, there are external socio linguistic factors - that is, social factors outside the language system. On the other hand, there are internal psycholinguistic ones - that is, linguistic and psychological factors which reside in the structure of the language and the minds of the speakers” (p. 134). These two categories are deeply intertwined. Every language system has underlying “weak points” due to the “internal” factors, and the “external” factors will accelerate linguistic changes by taking advantage of the vulnerable areas. Aitchison (2001) provides a metaphor for this argument: On closer examination, many [external sociolinguistic factors] turned out not to be 'real' causes, but simply accelerating agents which utilized and encouraged trends already existing in the language. When a gale blows down an elm tree, but leaves an oak standing, we do not believe that the gale alone caused the elm to fall. The gale merely advanced an event that would probably have occurred a few months or years later in any case. However, the gale dictated the direction in which the elm fell, which might in turn set off a further chain of events. If the elm fell against another tree. it might weaken this tree, and leave it vulnerable in another gale. Sociolinguistic causes of language change are similar to this gale. They exploit a weak point or potential imbalance in the system which might have been left unexploited. This exploitation may create further weak points in the system. (p. 151) 61 Of the “internal” elements, those involving phonology tend to cause the most dramatic linguistic changes. At its core, every spoken language is a system of coordinated sounds. In other words, sound itself is the building block of meaning. Every language has its own specific and finite set of sounds that, when combined, can create an infinite set of meanings. Although the difference between one single sound and another may seem insignificant, consider the distinction between the English words ‘hat’ and ‘hot’, or ‘nap’ and ‘map’. The smallest shifts in pronunciation produce an incredibly large semantic change. This principle is especially important in verb morphology, where a change in sound can lead to a change in a verb’s mood, aspect, tense, number, or person. For example, there is only a difference in one vowel between the phrases ‘I swim’ and ‘I swam’, but the tense has changed completely from present to past. Given the significance of sounds in words, it would be logical to assume that languages resist phonetic change. That assumption would not be entirely correct. If considered as a system of sounds, every language has areas in the system that are more susceptible to change than others. These “weak points” are incredibly common in all spoken languages; Aitchison’s Language Change: Progress or Decay? argues that “they are tendencies which are inevitably built into language because of the anatomical, physiological and psychological make-up of human beings” (Aitchison 2001: 153-154). Vulnerable spots in language include word-ending consonants, linking sounds by assimilation or omission, and articulatory effort in coordinating movements necessary to produce a sound sequence. Each of these “weak points” and still others have played an active role in the history of the Romance subjunctive. Word-final consonants, consonants at the end of a word that are not followed by a vowel, are the hallmark of the Classical Latin subjunctive person endings (m, s, t, mus, tis, nt). In 62 Vulgar Latin, the first person singular marker “m” had already been mostly eroded away. In Spanish, the third person singular “t” disappeared, and the third person plural “nt” has been reduced to “n” . In French, the third person singular “t” second person plural “tis” disappeared entirely, and the second person singular, first person plural, and third person plural final consonants are no longer pronounced. The changes in the subjunctive tense system can also be traced through the lense of linguistic ‘areas of vulnerability’. The original morphology of the Classical Latin imperfect subjunctive gradually fell into disuse due to confusion surrounding the shifting pronunciation of the word-medial “r”. The perfect and pluperfect subjunctive became permanently syncopated, wherein speakers omitted the “ve” or “vi” in the middle of the verb (which also acted as the perfective aspect marker). For example, the pluperfect subjunctive amāvisset was shortened to amāsset. This syncope eventually extended into writing, which complicated the situation of the perfect subjunctive. The syncopated form of the perfect subjunctive overlapped significantly with those of the future perfect indicative, as shown in Table 7 below. The only distinction between the two tenses was the first person singular, which is italicized. Table 8: Coinciding Morphology of the Latin Perfect Subjunctive and Future Perfect Indicative Syncopated Perfect Subjunctive Syncopated Future Perfect Indicative amārim amārō amāris amāris amārit amārit amārimus amārimus amāritis amāritis amārint amārint 63 Note: The discrepancy between the tenses in the first person singular is italicized. As perhaps can be expected in this scenario, the two tenses became frequently confused. Gradually, the future perfect indicative prevailed over the perfect subjunctive, and the perfect subjunctive morphology fell into disuse. Similar confusion between syncopated forms also occurred throughout the indicative mood. Eventually, the only original perfect tense still actively employed was the indicative preterite. But if the Latin imperfect and perfect subjunctive morphology no longer expressed the original tense “meaning”, how did the imperfect and perfect subjunctives survive into Modern Spanish and French hundreds of years later? The answer can be found by looking at the psychological aspects of language. Human beings have extraordinary memories, capable of retaining a seemingly infinite amount of information, especially in the realm of language. Indeed, humans have been capable of being polyglots for centuries. Although languages may seem like a tangled mess of words, they are actually highly organized structures. Aitchison (2001) states that: Every language contains a finite number of patterns [...] It is these patterns which enable humans to remember any language so apparently effortlessly. If the patterns were to break down, a person’s brain would become overloaded with fragmented pieces of information. Efficient communication would become difficult, if not impossible. (p. 169) Languages, or more accurately the speakers of languages, rely on patterns. These patterns are incredibly visible in verb morphology. Mood, tense, aspect, person, and number each have their own patterns that overlap to create meaning. The Latin verbal system, as well as those in French and Spanish, is created from binary divisions and symmetry. For example, the indicative mood and the subjunctive mood are opposed, and every tense in the subjunctive is linked derivationally to a tense in the indicative (present subjunctive and present indicative, imperfect subjunctive and 64 imperfect indicative, etc). The tenses are organized similarly, with each tense representing an intersection of aspect and time; Latin has three major timeframes (past, present, future) and two major aspects (imperfective and perfective), which combine into six tenses in the indicative mood. The subjunctive mood only has two major timeframes (past and present) and two major aspects (imperfective and perfective), for a total of four tenses (See Table 1). Each tense has its own pattern (the morphological endings). It can therefore be said that the Romance languages do not change from Latin with regard to the morphological typology of the language. When linguistic organization breaks down, languages apply therapeutic changes to maintain the patterns which convey overall meaning. This can be seen in the development of the Latin subjunctive; when the future perfect indicative and the perfect subjunctive syncopated forms became too similar to distinguish, the perfect subjunctive fell out of use, preserving the one form/one meaning principle that a given grammatical element must only correspond to a singular interpretation. Further disruption from the disintegration of the imperfect and pluperfect subjunctive tenses threatened the symmetrical arrangement of the subjunctive system. Several changes occurred over time to repair the mood. The original forms of the pluperfect subjunctive (combined with the also defunct pluperfect indicative in the case of Spanish) were reanalyzed and recategorized as the imperfect subjunctive tense, reestablishing the two-tense imperfective aspect in the subjunctive. The economy principle in linguistics asserts that language development moves generally toward the ease of implementation and interpretation; that which unnecessarily complicates communication will be gradually eliminated. This accessibility of significance-- that is, how efficiently and effortlessly meaning can be conveyed-- is a driving force in the linguistic evolution of the Romance subjunctive. When morphological forms became unclear, entered into 65 competition with one another, or fell out of use entirely (as seen in the imperfect, perfect, and pluperfect Latin subjunctives) change was triggered to conserve the morphological patterns upon which the subjunctive was designed. The patterns that organized the subjunctive system were essential to effective communication, and established meaning by coordinating aspect, tense, person, and number morphology. Although the system may appear complicated, in fact it conserved the accessibility of significance with its rigid structure. The economy principle, then, moves to protect this system, with the longstanding goal of increasing comprehensibility. The economy principle is also observed in the survival of the perfect system in both the indicative and subjunctive moods through the propagation of the periphrastic perfect. As previously discussed, this construction of the auxiliary verb habēre “to have” and the past participle of a verb did already exist in Classical Latin, but was rarely used. Over time, the construction became more popular (and semantically reinterpreted), eventually competing with the indicative preterite. Once the periphrastic form of the present perfect had been established, the past perfect was quick to follow, a process called analogical change. After all, the patterns dictating the verb system relied on symmetry; a singular tense construction (like “I have done”) could not exist for long without being extrapolated to apply to other tenses (like “I had done”). Changing the tense of the new periphrastic perfect was only a matter of conjugating the auxiliary verb. Eventually, the periphrastic perfect restored the “missing” perfect and pluperfect subjunctive. In conclusion, language change in general and the development of the Romance subjunctive in particular cannot be reduced to laziness or simplification, nor can it be reduced to random interference of entropy. Although linguistic evolution is not completely predictable, 66 there are certain identifiable points that are susceptible to change. Certain elements of language are stable, others variable. Some languages have developed more extensively than others, and even within a given language, evolutions may happen in only one dialect or region, while others remain unchanged. “External” social factors, discussed in depth in the next chapter, play a large role in determining which weaknesses will be exploited and thus undergo change. Whether it is sparked by internal or external forces, any shift that disrupts the patterns which govern language will prompt correction in order to maintain linguistic coherence. 67 “‘Contrariwise,’ continued Tweedledee, ‘if it was so, it might be; and if it were so, it would be; but as it isn’t, it ain’t. That’s logic.’”- Through the Looking Glass by Lewis Carroll 3. THE PRAGMATIC EVOLUTION OF THE SUBJUNCTIVE MOOD: FROM CLASSICAL LATIN TO MODERN ROMANCE Now that the development of the morphological forms themselves has been cataloged in chapter 2, this next chapter will detail the pragmatic evolution of the subjunctive mood, that is, how the semantic use of the subjunctive mood has changed over time from Classical Latin through Modern Spanish and French. The true “meaning” of the subjunctive mood has always been a polemic issue; some scholars argue that employing the subjunctive is a “meaningless” mechanical gesture, a grammatical relic for marking subordinate clauses. Still others insist that the subjunctive mood has developed a deep psychological richness over time, and the use of the subjunctive mood can signify conscious and unconscious relationships to truth and value. Meaningful or meaningless, the odyssey of the subjunctive mood continues through the centuries, and its progress, as well as an analysis through the lens of linguistic theory, is discussed in this chapter. 3.1 Latin While linguistics and historians alike have been investigating the Latin language for hundreds of years, there has been little academic consensus on the subjunctive mood. In fact, grammatical studies of the Latin subjunctive have been quite contentious, and scholars remain divided on key issues even today. Intellectual debate begins chronologically with the mere existence of the 68 subjunctive in Latin. Due to its complex and multifaceted nature compared to other Indo-European languages, earlier linguistic theorists argued that the Latin subjunctive system was inherited from the Proto-Indo-European (PIE) subjunctive mood and then expanded upon (cf. Exon 1922, Ferrell 1999). Modern researchers of Latin and PIE linguists, however, contend that the intricacies of the Latin subjunctive arise from its origin as a combination of two Proto-Indo-European moods: the subjunctive, used for hypothetical events or events in the remote future, and the optative, used for expressing wishes or hopes (cf. Ferrell 1999). Interestingly, although the pragmatic functions of the two PIE moods were fused together into the Latin subjunctive, contemporary researchers attest that the morphological forms did not merge; while the verb endings of the PIE optative continued into the Latin subjunctive, the PIE subjunctive did not. Indo-European Linguistics: An Introduction, Clackson states that “in Latin, for example, inherited subjunctive forms are continued as futures, and inherited optative forms are continued as subjunctives” (Clackson & Meissner 2007: 120). Even after setting questions of ancestry aside, the Latin subjunctive continues to perplex linguists with its diversity of semantic uses. The following sections will attempt to illustrate the variety of linguistic contexts in which the Latin subjunctive appears, but an exhaustive list is beyond the scope of this study, as is a true inspection and analysis of Latin sources. Instead, this study will rely on the research of others, especially Murphy (2008), as a basis for investigation. 3.1.1 Classical Latin By the beginning of the Classical Latin period, the presence of the subjunctive mood had been firmly established in the writings of Roman grammarians and rhetoricians. Precise rules and interpretations for the subjunctive, however, proved to be elusive. Scholars of every area have 69 tried to squeeze the subjunctive system into a narrow set of rules or to stretch the rules in order to account for the subjunctive system, but efforts have been mostly in vain. This study will be the next in a long line of similar works to attempt to catalog the numerous linguistic contexts of the Classical Latin subjunctive. Using frameworks from Murphy (2008), Allen & Greenough (2014), and Batstone (2022), the subjunctive system is first broken into three syntactic sections: main (independent) clauses, subordinate (dependent) clauses, and indirect discourse. Each of these sections is further subdivided by pragmatic use. 3.1.1.1 The Subjunctive in Main Clauses The main clause of a sentence, also called the independent clause, is the part of the sentence that can form a complete sentence ‘by itself’, i.e. has a grammatical subject and predicate. The Classical Latin subjunctive was frequently used in independent clauses. Batstone (2022) identifies three broad categories of subjunctive use in the main clause: the jussive subjunctive, the potential subjunctive, and the optative subjunctive. In this organization, the jussive can express direct commands, or ask a question if the expected answer is a command; unlike the imperative mood, it can be employed in the present or the past, and can be used in any verbal number or person. The jussive subjunctive is “the direct expression of the duty or obligation that rests on someone” (Batstone 2022: section 2). The potential subjunctive expresses “an action as existing in the realm of possibility”, and may be used for hypotheticals, “polite hesitation”, or making a request “regarding how the addressee feels about the potential of likelihood of an event” (Batstone 2022: section 3). Batstone categorizes the optative subjunctive as expressing “an action as existing as part of the speaker’s wishes, prayers, or fears [...] a direct expression of emotional desire” (section 4). 70 Murphy (2008) similarly classifies the independent uses of the Classical Latin subjunctive into the three categories of optative, volitive (an equivalent of Batstone’s jussive), and potential. The volitive category is additionally divided into five parts: the jussive for commands; the hortatory for first person plural commands and exclamations; the prohibitive for prohibitions; the deliberative, which Batstone (2022) explicitly refutes, for deliberating exclamations; and the concessive for granting something. The potential subjunctive is divided into four additional parts: “may” for mere possibilities; “would” for understood conditions; “could” for perceiving and thinking; and the “should/would/could have” for irrealis. This framework, along with the corresponding Classical Latin examples and their English translations from Murphy (2008), is presented in Table 8 below. Table 9: Uses of the Subjunctive in Classical Latin Independent Clauses Optative Subjunctive Volitive Subjunctive Potential Subjunctive Optative di istaec prohibeant may the gods prevent that Jussive dicant let them tell “May” dicat aliquis someone may say Hortatory amemus patriam let us love our country “Would” dies deficiat time would fail Prohibitive ne repugnetis do not resist “Could” crederes one could believe Deliberative quid faciam what shall I do “Should/Would/Could Have” fuissem philosophus I would have been a philosopher 71 Concessive sit hoc verum I grant that this is true Note: The subjunctive conjugations are bolded. 3.1.1.2 The Subjunctive in Subordinate Clauses Classical Latin also makes use of the subjunctive in subordinate clauses, also called dependent clauses, which cannot ‘stand alone’ as a complete sentence but instead provide additional information to support the main clause. Subordinate clauses can be divided by grammatical function (either as substantive, adjectival, or adverbial), but here they will continue to be classified by pragmatic function using the analyses of Batstone (2022) and Murphy (2008). Batstone (2022) claims that the uses of the subjunctive in dependent clauses “have developed from the three original uses of the independent subjunctive”, and can thus be divided by the same three categories of the optative, jussive, and potential (section 5). Murphy (2008), on the other hand, designates thirteen types of subjunctive use in dependent clauses. Each of these thirteen are mentioned by Batstone (2022). The single discrepancy between the two frameworks is the fear clause, which Batstone (2022) names as the only Classical Latin optative subjunctive in subordinate clauses. Murphy (2008) does not reference fear clauses at all, so in assimilating the two schemes liberty was taken to assign fear clauses as a type of purpose clause. The (now fourteen) established subordinate clauses in which the subjunctive is used are not divided equally amongst the three broad subjunctive categories. The optative, the broad category for wishes or desires, possesses only fear clauses, which express a fear that something will (preceded by the conjunction ne) or will not (preceded by the conjunction ut) happen. The 72 volitive (Batstone’s jussive) has only two clause types: final (purpose) clauses, which express an event that occurs in order that or lest something else happen, and proviso causes expressing an event occurring provided that another action happens. Semantically, the volitive category unites these two clause types under the reasoning that it “gives direct expression to someone’s will, purpose, goal, and so on” (Batstone 2022: section 6). By far the most extensive categorization for the use of the subjunctive in subordinate clauses is the potential, which according to Batstone (2022) “is used to make statements in which the speaker represents an event as something he or she has imagined or thought about as likely or possible” (section 7). Consecutive (result) clauses describe an event that occurs so that or that [not] another event or state occurs. Adversative and Concessive clauses express an event that happened although a seemingly contradictory event also occurred. Causal clauses express the cause of an event, usually with the conjunction cum ‘since’. Similarly, circumstantial “cum” clauses express relevant circumstances pertaining to the main clause using cum. Clauses of characteristic are clauses that provide additional information or describe the main clause. “Anticipatory” temporal clauses express an event that occurs until a second event occurs. The potential subjunctive also includes conditional clauses; the protasis (the second part of the condition) of both “ideal” and “unreal” conditions, usually seen with si ‘if’, as well as comparison conditional clauses, which usually are translated with the English phrase as if or as though. Table 9 below illustrates the fourteen clause types, as well as the general subjunctive category to which they belong. 73 Table 10: Uses of the Subjunctive in Classical Latin Dependent Clauses Optative Subjunctive Volitive Subjunctive Potential Subjunctive Fear (Purpose) vereor ut veniat I’m afraid that he won’t come Final (Purpose) ut omnes intellegant dico I say it so that everyone understands Consecutive (Result) Siciliam ita vastavit ut resituti in antiquum statum non possit he so ravaged Sicily that it cannot be restored to its former condition Adversative & Concessive licet omnes terrores impendeant, succurram although all terrors hang over me, I will lend aid Characteristic multa sunt, quae mentem acuant there are many things which sharpen the wits Causal cum sis mortalis, quae mortalia sunt cura since you are mortal, care for what is mortal Proviso nubant, dum ne dos fiat comes let them marry, provided that no dowry goes with it Circumstantial “Cum” cum ver appetat, ex hibernis movendum est as spring is approaching, we must move out of our winter quarters “Anticipatory” (Temporal) dum litteras veniant, morabur I shall wait until the letter comes Comparison (Conditional) 74 serviam tibi tam quasi emeris me argento I will serve you as though you had bought me for money Protasis of “Ideal” (Conditional) si hoc dicas, erres if you should say this, you’d be mistaken Protasis of “Unreal” (Conditional) sapientia non expeteretur, si nihil efficeret wisdom would not be sought, if it accomplished nothing 3.1.1.3 The Subjunctive in Indirect Discourse The last syntactic context where Classical Latin employs the subjunctive is in indirect discourse: indirect statements and indirect questions. These two types of subordinate clauses, also referred to as reported speech, are both classified by Batstone as belonging to the jussive (known as the volitive in this paper) subjunctive category. Batsone (2022) presents a straightforward rule for employing the subjunctive in indirect discourse: The rule in Latin is actually simpler that is sometimes presented: If a clause was in the original statement, it must be placed in the subjunctive when the original statement is reported [...] This rule has certain logical consequences. First, if a clause in the middle of a report in indirect discourse has a verb in the indicative, this means only that the speaker is adding this point at the time that he is speaking, and that he is not claiming that they were part of the original statement. There is one apparent exception to this: that is, when both the original statement and the speaker could be referring to a clear and present fact, then the speaker often puts that fact in the indicative -- but this usually applies to things at hand, things seen or things the speaker could point to. Second, if a clause is in the subjunctive, this does NOT mean that the speaker disagrees or reserves judgment. [...] Furthermore, it is an observable fact that a speaker who reports his own thinking about a subject distinguishes between what he was thinking in the past (which will go into the subjunctive) and what his current reasoning is (which will have to be in the indicative). (section 9) 75 3.1.1.4 The Sequence of Tenses The consecutio temporum, also known as the sequence of tenses, is a set of grammatical rules which dictate the use of subjunctive tenses within a subordinate clause. In order to have chronological harmony within the sentence and to preserve nuance in semantic meaning, the tense of the subjunctive in the subordinate clause is governed by the tense of the verb in the main clause. The majority of independent-dependent structures in Latin used the main verb in the indicative mood in the independent clause, and the secondary verb in the subjunctive mood in the dependent clause. As described in chapter 2, the Classical Latin indicative mood is a paradigm that is symmetrical on two axes: time of reference and aspect. There are three time frames (past, present, and future) and two aspects (incomplete/imperfective and complete/perfective), which combine into a total of six indicative tenses: present, perfect, imperfect, pluperfect, future, and future perfect. The subjunctive mood, however, has only four tenses (there is no future time, and thus no future or future perfect tenses). In the sequence of tenses, the indicative tenses are split into two sequences. The main clause is in a primary (also called “non-past”) sequence when the main verb is conjugated in the present, future, or future perfect tenses. The main clause is in a secondary (also called “past”) sequence when the main verb is conjugated in the imperfect, perfect, or pluperfect tenses. To determine the tense of the subjunctive verb in the subordinate clause, the aspect of the main verb must be taken into account. In a primary sequence main clause, the subjunctive verb in the subordinate clause will be conjugated to the present tense if the aspect of the main indicative verb is incomplete, or to the perfect tense if the aspect of the main verb is complete. In a 76 secondary sequence main clause, the subjunctive verb in the subordinate clause will be conjugated to the imperfect tense if the aspect of the main indicative verb is incomplete, or to the pluperfect tense if the aspect of the main verb is complete. Of course, as with any good grammatical rule, there are always exceptions. The “exception” most relevant to this present study is that after a primary sequence, the perfect subjunctive is often used in the subordinate clause to indicate any event in the past, regardless of aspect. 3.1.2 Vulgar Latin Compared to the extensive research conducted on the subjunctive in Classical Latin, there seems to be precious little information readily available on the subjunctive in Vulgar Latin, presumably due to the difficulty locating reliable and relevant sources. Most conclusions about the development of the subjunctive in the Vulgar Latin period are retrospective, and rely on early Romance varieties to bridge the gap between Classical Latin and the modern Romance family. Even without a large-scale corpus of Vulgar Latin, however, an analysis of the subjunctive is still possible because of the number of salient developments occurring between Classical Latin and Old Romance languages. The common academic view today is that beginning already in Late Classical Latin, linguistic evolution of the subjunctive had begun, and change continued throughout Vulgar Latin. In fact, Grandgent (1907) notes that “at the end of the Vulgar Latin period [the subjunctive] was probably used, in popular speech, very much as it is used in the Romance languages” (p. 52). 77 3.1.2.1 The Subjunctive in Main Clauses Shifts in the pragmatic sphere of the Vulgar Latin subjunctive seem to have started with the use of the subjunctive in independent clauses. While in Classical Latin, the subjunctive was frequent in main clauses-- Murphy (2008) alone identifies ten distinct uses-- in Vulgar Latin the prevalence of the ‘independent’ subjunctive decreased significantly. Although hypothetically any verb in the language could appear in the subjunctive, Vulgar Latin sees a gradual reduction in the variety of expressions, almost as if speakers were reticent to branch out from the most common phrases ingrained in common speech. Grandgent (1907) also comments on this phenomenon, observing that “late writers, while trying to follow the traditional practice, were less logical and evidently less spontaneous than Classic authors in their employment of the subjunctive” (p. 53). 3.1.2.2 The Subjunctive in Subordinate Clauses A fascinating contradiction in Vulgar Latin is that although a cursory examination of the subjunctive conventions might (and did) lead to the determination that the Vulgar Latin subjunctive mood was withering away, upon closer analysis researchers have discovered that innovation was indeed occurring within the mood. Overall, use of the subjunctive in subordinate clauses did diminish, but the subjunctive also began to appear in subordinate clauses where in Classical Latin it was not required. In other words: “the subjunctive was limited to fewer constructions, being replaced by the indicative in many constructions [...] On the other hand, late writers often put the subjunctive where Classic authors would have put the indicative” (Grandgent 1907: 52-53). Specifically, the indicative began to encroach upon conditions that were not contrary to fact (i.e., the protasis of “ideal” conditional clauses and comparison 78 conditional clauses in Table 7 above), as well as all dependent clauses except dubitative or adversative. Only in the modern era of research has an academic consensus been reached on the linguistic reasoning behind these developments. Comprehensive cross-linguistic grammatical analyses, notably Murphy (2008) and Digesto (2019), support the now widely accepted theory that the rise in subjunctive dependent clauses is originally syntactically, rather than semantically, motivated. Digesto (2019) presents research on the shift from paratactic to hypotactic sentence structure. In parataxis (which can be translated as “arranging side by side”), sentences are grammatically simpler, with independent clauses standing alone or joined by coordinating conjunctions. In contrast, hypotaxis (which can be translated as “arranging under”) uses subordinating conjunctions to connect main clauses with subordinate ones. Classical Latin rhetoric favored parataxis, which explains the pervasive use of the subjunctive in independent clauses. Vulgar Latin preferred hypotaxis, which justifies both the waning of independent subjunctives and the expansion of the subjunctive in dependent clauses. This change is delineated in Table 10 below. Table 11: Latin Protactic to Hypotactic Shift Protaxis Hypotaxis volo; veniat I wish it; let him come volo ut veniat I want him to come [that he come] Note: Subjunctive conjugations are bolded. Adapted from Digesto (2019: 203). Murphy (2008) takes this semantic investigation one step further, and determines that the transition from protaxis to hypotaxis is part of a larger shift from left-branching sentences, in 79 which the subordinate clause comes before the main clause, to right-branching sentences, in which the subordinate clause follows the main clause. Moving from left- to right-branching structures led to an increase not only in the frequency of subordinate clause use, but also in the use of subordinating conjunctions that would mark the link of the independent and dependent clauses. In Classical Latin, subordinate clauses were marked by relative pronouns, which would decline in accordance with the Latin case system to agree in gender, number, and case with the element being modified. This concept may seem unusual to the contemporary English speaker, but consider the subtle difference between “the books which are on the shelf” and “the girl who walked to the store”; a native speaker would never say “the books who are on the shelf”, as “who” is reserved for people. Further nuance in the English system occurs with the now antiquated pronoun “whom”, which can only be used as a direct object: grammatically speaking, a letter must always be addressed with the phrase “to whom it may concern”, not “to who it may concern”, because the relative pronoun is the object of the phrase, not the subject. An analogous, but much more extensive, system of relative pronouns was used in Classical Latin. Declining relative pronouns continued into Vulgar Latin, but gradually the complex framework grew simplified. A set of five accusative (direct object) forms came to be used almost exclusively to begin subordinate clauses: quom, quando, quoniam, quia, and quod. Each of these conjunctions originally retained a pragmatic significance (which can perhaps be compared to the difference in English between the relative pronouns “since”, “when”, “which”, “who”, and “that”). Eventually, nevertheless, over the course of expanded use the semantic meaning of these forms weakened. Three conjunctions (quoniam, quia, and quod) evolved into “empty” complementizers, and were decategorized completely. In the later era of Vulgar Latin, this 80 development culminated in a specialization process, wherein the three complementizers competed for use. Quod emerged as the most selected conjunction (Murphy 2008: 282). These modifications to the relative pronoun systems, combined with the generalization of the subordinating conjunctions ut and cum, were intensified by the transition from left-branching/protactic sentences to right-branching/hypotactic ones. The final denouement of this process was that the subjunctive mood became increasingly associated with subordinate clauses, rather than main ones. Poplack et al. (2018) succinctly summarizes one perspective on centuries-long grammaticalization of the subjunctive mood: The subjunctive is first used in complement clauses with the same meaning it has in main clauses. As such uses increase in frequency, the semantic contribution of the subordinate mood weakens (presumably abetted by the redundancy of expressing modality twice), eventually leading to its (re-)analysis as a concomitant, if not a marker, of subordination. Once this has occurred, the subjunctive may spread to embedded clause types whose meaning is no longer harmonic with that of the matrix clause, and where it itself makes little or no semantic contribution. After this, the only development left is extinction. (p. 5) However, as mentioned above, in Vulgar Latin the subjunctive mood is quite obviously not extinct, nor is it in danger of becoming so, but elements of the scenario outlined above do appear congruent with current understandings of the system of the era. Digesto (2019) suggests that “syntactic reorganization, and the recognition of subjunctive morphology as a marker of subordination, would have led to the gradual permeation of the subjunctive into syntactic environments that were traditionally associated with the indicative” (p. 203). This is roughly where the story of Vulgar Latin leaves the subjunctive mood. 81 3.1.2.3 The Subjunctive in Indirect Discourse In both indirect statements and indirect questions, the substitution of the indicative for the subjunctive was common in Vulgar Latin, leading some theorists to argue that the shift had been completed by 200 BCE (cf. Digesto 2019). 3.1.2.4 The Sequence of Tenses The sequence of tenses experienced shifts in Vulgar Latin that would ultimately become significant for the Romance languages. Herman (2000) remarks that the consecutio temporum was a “fairly strict rule” in Classical Latin that “required a clear idea of the nature of tenses and the precise use of quite a subtle grammatical device, so it is perhaps only to be expected that the rule is relaxed in the Vulgar Latin texts” (p. 93). The most pertinent change in the sequence of tenses is that, in a secondary (“past”) sequence main clause, Vulgar Latin regularly substituted the pluperfect subjunctive for the imperfect subjunctive. This caused the pluperfect subjunctive to lose its perfect aspectual connotation over time, which is a contributing factor to the gradual morphological replacement of the imperfect subjunctive by the non-periphrastic (one-word) pluperfect forms. 3.2 Spanish Compared to other Romance languages, Spanish is generally upheld as a more “conservative” language (cf. Lindschouw 2010, Poplack et al. 2018). In other words, although every member of the Romance language family originated from Latin, the Spanish of today remains relatively close to Latin from a grammatical perspective. That is not to say that Spanish is a repackaged version of Latin: over the centuries, Spanish has developed a unique linguistic identity separate 82 from Classical and Vulgar Latin. Spanish has also experienced major shifts in grammar, especially in the realm of the subjunctive mood. The following sections attempt to document some of the changes that occurred in the odyssey of the subjunctive through Vulgar Latin into Modern Spanish. 3.2.1 Old Spanish If Modern Spanish is categorized as a “conservative” language, it may perhaps be accurate to describe Old Spanish as an ‘experimental’ one. It is during the Old Spanish period that several significant grammatical elements developed. Some changes endured through the present day, others had a lifespan similar to that of Old Spanish itself. The subjunctive mood evolves both morphologically and pragmatically, and many shifts are so interconnected that it is difficult to separate the changes in form from changes in meaning. These changes may appear superficial, but they are much deeper than surface adjustments, and have ramifications that will continue to shape the Spanish subjunctive system into the future. 3.2.1.1 The Subjunctive in Main Clauses In main clauses, the subjunctive mood appears less frequently in Old Spanish than it did in Classical Latin, but more frequently than in Modern Spanish. Of the three broad categories of subjunctive use, the optative and volitive closely resemble those in Latin. In main clauses, Old Spanish selects the subjunctive in independent clauses to express wishes (optative subjunctive), commands (jussive and hortatory subjunctive), prohibition (prohibitive subjunctive), and concession (concessive subjunctive). In the concessive constructions, pragmatic meaning extends to permission or acceptance, polite reservation, or even “resignation in front of inevitable circumstances” (Jensen & Lathrop 2017: 19). However, notes The Syntax of the Old Spanish 83 Subjunctive, “this usage seems restricted to certain set formulas which mainly show repetition of the same verb” (Jensen & Lathrop 2017: 19). Jensen & Lathrop (2017) also identify a new category of the subjunctive: the subjunctive of doubt, which in this present paper will be assimilated into the potential category. The subjunctive of doubt emphasizes uncertainty, and is usually found with adverbs like quizá, por ventura, and acaso. The original independent uses of the potential subjunctive category are not mentioned in The Syntax of the Old Spanish Subjunctive. 3.2.1.2 The Subjunctive in Subordinate Clauses As in Vulgar Latin, there is an extension of subjunctive use in subordinate clauses in Old Spanish. The Vulgar Latin specialized complementizer quod was gradually overtaken by the early Romance hybrid form que, which Old Spanish inherited. Murphy (2008: 282) explains that “finally, there was a process of renewal, whereby the restructuring of the entire conjunction system took place, building new forms based on the highly-grammaticalized Romance element que (e.g. Sp. porque, para que, aunque, después de que…)”. With the new conjunction system, Old Spanish continued to select the subjunctive in many of the same dependent clause types as Latin, including final purpose clauses, consecutive result clauses, and adversitive/concessive clauses. Yet a clear shift occurred in the logic of subjunctive selection: not only had the subjunctive mood become associated with subordinate clauses, more importantly, it became explicitly associated with doubt and uncertainty. Use of the indicative signifies that an eventuality in a clause either already happened or is seen as objective fact, while use of the subjunctive expresses possible eventualities or subjectivity. Consequently, it is no surprise that 84 Jensen & Lathrop (2017:42) remark that “volition constitutes the strongest area of the use of the subjunctive. The subjunctive is very firmly entrenched here, allowing the indicative to appear only when the volitive element is weakened”. A related factor in mood selection that becomes increasingly critical in Old Spanish is the concept of “governance”. Governing verbs, verbs that appear in the main clause of the sentence, determine the modality of the subordinate clause. Jensen and Lathrop (2017:57-58) note: [...] the dividing line between the use of the indicative or the subjunctive is determined by the degree of certainty or doubt [in the governing verb]. The indicative appears with a high degree of certainty, whereas actions which are only possible, doubtful, or impossible are expressed in the subjunctive mood. With expressions centered around this dividing line in modal usage [...] modal fluctuation is not uncommon. Certain governing verbs exclusively select the indicative while others (those expressing will, wishes, desires, commands, pleas, endeavors, prevention, necessity, judgment etc.) exclusively select the subjunctive. Yet another class of verbs and verbal phrases (including ver “to see”, saber “to know”, decir “to say”, and ser cierto “to be certain/correct”) govern the indicative when positive, and the subjunctive when negative. For example, saber “to know” requires the indicative mood, but no saber “to not know” introduces uncertainty and thus requires the subjunctive. Verbs of fearing take the subjunctive tense in connection with future eventualities that are uncertain. This use of the subjunctive can be traced back to the Latin optative subjunctive category. Verbs of astonishment also select the subjunctive, which according to Jensen & Lathrop (2017) is due to the pragmatic similarity between astonishment and uncertainty. Confusingly, other emotion verbs, even “verbs of high emotional content”, take the indicative. These tendencies fluctuate throughout the thirteenth through the sixteenth centuries. Surprisingly, 85 there was a relatively brief rise in usage of the indicative with fear clauses. Ferrell (1999) postulates that the disorientation is rooted in the creation process of the future and conditional tenses. Appearances of the indicative in fear clauses occurred mainly with verbs conjugated into the future and conditional tenses, which were still undergoing grammaticalization at the time; once these tenses became firmly established as non-subjunctive, the use of the indicative in fear clauses decreased. Another complication in expressions of emotion is that early Old Spanish prefers the subjunctive for expressions of pleasure, and the indicative for expressions of hope and all other emotions. By late Old Spanish, however, the subjunctive came to be used with all verbs of emotion. Ferrell (1999:199-200) concludes that: Expressions of pleasure and hope convey the same type of nuance carried by verbs such as querer ‘to want’ or desear ‘to desire’ that always take the subjunctive in accompanying noun clauses. The use of the subjunctive with these expressions of emotion most likely came about by analogy to expressions of will or desire. As to why the subjunctive came to be used with all the other expressions of emotion, matters are not so clear. It is possible that, as in Latin, speakers wanted to convey a nuance of emphasis or disassociation, which was possible to convey only by use of the subjunctive. It is also possible that these other expressions of emotion followed an analogical pattern set by expressions of pleasure and hope, and began to use the subjunctive by analogy to them. It appears that this use of the subjunctive with expressions of emotion in general was a strong tendency, at least in the literary language, by the mid-sixteenth century, and appears to have taken place sometime during the period between the late fifteenth century text La Celestina, and the translation of Cicero’s Epistolae ad familiares in the sixteenth. To summarize the numerous developments in the use of the Old Spanish subjunctive in subordinate clauses is to illustrate a large-scale process of association and assimilation. Aided by the uses of the optative, volitive, and potential Latin subjunctive categories, Old Spanish gradually links subjunctive selection almost exclusively with the concepts of wish and uncertainty. A semantic binary, unclear at times, is constructed between the indicative-- representing undisputed fact, certainty, and objectivity-- and the subjunctive-- representing 86 possibility, uncertainty, and subjectivity. Main clause verbs become the most important factor in dictating the mood of subordinate clauses. Over the period of Old Spanish, mood selection is eventually completely reorganized to be consistent with the indicative-subjunctive distinction. 3.2.1.3 The Subjunctive in Indirect Discourse In accordance with the informal rules outlined above, indirect questions and statements selected the subjunctive only when expressing doubt (often with a negated governing verb), obligation, purpose, or intention. In other cases, the indicative was used. 3.2.2 Modern Spanish The pragmatics of the Modern Spanish subjunctive system are quite complementary to those of the (late) Old Spanish one. The developments inherited from Old Spanish (which in turn were inherited from Vulgar Latin, and so on) continue to progress through Modern Spanish. 3.2.2.1 The Subjunctive in Main Clauses The following table, adapted from Murphy (2008), is the same as Table 8 above, and presents the same subjunctive categories. Each original use of the subjunctive is color coded; the lighter the color, the stronger the subjunctive selection bias. Boxes with a white background require a subjunctive almost exclusively. Boxes with a light gray background take the subjunctive, but also accept the indicative. Boxes with a black background have been replaced entirely by a non-subjunctive mood. Table 12: Uses of the Subjunctive in Modern Spanish Independent Clauses Optative Subjunctive Volitive Subjunctive Potential Subjunctive Optative que los dioses lo impidan may the gods prevent that Jussive que lo digan let them tell “May” alguien podría decir someone may say 87 Hortatory amemos nuestro país let us love our country “Would” el tiempo fallaría time would fail Prohibitive no resistan do not resist “Could” uno podría creer one could believe Deliberative que debería hacer what shall I do “Should/Would/Could Have” habría sido filósopho I would have been a philosopher Concessive concedo que eso es verdad I grant that this is true Note: The subjunctive conjugations are bolded. Out of the original ten uses of the subjunctive in independent clauses employed in Classical Latin, only four remain restricted to the subjunctive mood in Modern Spanish: the optative, which expresses hopes, wishes, fears, and desires; the jussive and hortatory, which express commands for the third persons and the first person plural respectively; and the prohibitive, which expresses a prohibition or a negative command. Two main clause uses accept either the subjunctive or a non-subjunctive mood (the indicative or the conditional): the concessive, which expresses permission or concession; and the “may” potential, which expresses mere possibility. The four remaining categories can be divided by their volitive and potential classifications. The deliberative, the only volitive use to no longer accept the subjunctive, has been replaced by a range of modal options, including the indicative, the infinitive, or the 88 conditional mood. The three potentials-- “would”, “could”, and “should/would/could have”-- exclusively select the conditional mood. Also of note is the prevalence of que in front of the optative and jussive subjunctives, even though both occur in independent clauses. This phenomenon evidences the identification of the subjunctive mood with subordination (i.e., relative pronouns that subordinate clauses). Except for prohibitive uses of the subjunctive and “stock phrases” like viva España! (“long live Spain!), the majority of subjunctive phrases in Modern Spanish are preceded by que or another subordinating conjunction. 3.2.2.2 The Subjunctive in Subordinate Clauses The following table, adapted from Murphy (2008), similarly to Table 8 above, and presents the same subjunctive categories. Each original use of the subjunctive is color coded; the lighter the color, the stronger the subjunctive selection bias. Cells with a white background require a subjunctive almost exclusively. Cells with a light gray background take the subjunctive, but also accept a non-subjunctive construction (e.g. the indicative, the infinitive, the conditional). Cells with a black background have been replaced entirely by a non-subjunctive mood. Table 13: Uses of the Subjunctive in Modern Spanish Dependent Clauses Optative Subjunctive Volitive Subjunctive Potential Subjunctive Fear (Purpose) temo que no venga I’m afraid that he won’t come Final (Purpose) lo digo para que todos entiendan I say it so that everyone understands Consecutive (Result) destrozó Sicilia de tal manera que no puede ser restaurada a su condición original he so ravaged Sicily that it cannot be restored to its former condition Adversative & Concessive aunque todos los miedos me 89 acechen, prestaré ayuda although all terrors hang over me, I will lend aid Characteristic hay muchas cosas que agudizan el ingenio there are many things which sharpen the wits Causal ya que eres mortal, preocúpate por lo mortal since you are mortal, care for what is mortal Proviso que se casen, siempre que no haya dote let them marry, provided that no dowry goes with it Circumstantial “Cum” como se acerca el invierno, debemos trasladarnos de nuestro campamiento de invierno as spring is approaching, we must move out of our winter quarters “Anticipatory” (Temporal) esperaré a que llegue la carta I shall wait until the letter comes Comparison (Conditional) te serviré como si me hubieras comprado por dinero I will serve you as though you had bought me for money Protasis of “Ideal” (Conditional) si dijeras eso, estarías equivocado if you should say this, you’d be mistaken Protasis of “Unreal” (Conditional) 90 la sabiduría no sería buscado si no consiguiera nada wisdom would not be sought, if it accomplished nothing Note: The subjunctive conjugations are bolded. Within dependent clauses, the subjunctive still shows a selection bias, presumably because of the strong association between the subjunctive mood and subordinate clauses. Five of fourteen original Latin categories solely take the subjunctive: the optative, for clauses expressing fear and other emotion; proviso clauses expressing events occurring provided that another event occur; as well as three uses of the potential subjunctive classification. These conditional clauses are linked to the subordinating conjunction si “if” (como si “as if” in the case of the comparison conditional clauses). An equal number of categories accept either the subjunctive or a non-subjunctive mood: purpose clauses (which prefer the subjunctive and are usually marked with the subordinating conjunction para que); consecutive (result) clauses; adversitive/concessive clauses (which prefer the indicative, but must take the subjunctive after the subordinating conjunction aunque); clauses of characteristic; and “anticipatory” temporal clauses. Much like in Old Spanish, the variation in mood selection depends on the level of certainty: in clauses of characteristic and “anticipatory” temporal clauses especially, the indicative mood is used to denote a known or definite eventuality, while the subjunctive is used for unknown or indefinite ones. For example, in the “anticipatory” temporal clause sentence given in the table, the verb is in the subjunctive. If the verb were in the indicative mood, it would indicate that the speaker knows exactly what time the 91 letter will arrive and has a definite time in mind. Since the verb is in the subjunctive, it indicates that the speaker does not know the precise time of the letter’s arrival. Only two categories of the subjunctive in subordinate clauses have transitioned completely to indicative selecting: causal clauses and circumstantial clauses. This development is also explained by the pragmatic contrast between indicative and subjunctive. Whereas in Latin these clauses would only take the subjunctive, usually with the conjunction cum, in Spanish phrases like ya que eres mortal “since you are mortal” and como se acerca la primavera “as [since] spring is approaching” are interpreted as statements of objective fact (i.e., you are indeed mortal, spring is indeed approaching) and thus select the indicative mood. 3.3 French French is considered by linguists to be one of, if not the most, “innovative” languages in the Romance family (cf. Lindschouw 2010, Poplack et al. 2018). In other words, from a grammatical perspective, Modern French has evolved relatively far from Classical Latin. This is especially true in the realm of the subjunctive. The following sections attempt to document some of the changes that occurred in the odyssey of the subjunctive through Vulgar Latin into Modern French. 3.3.1 Old French Despite the “innovative” characterization of Modern French, Old French retains a subjunctive system that is quite similar to that of Vulgar Latin. Indeed, it could be argued that Old French is less ‘experimental’ than Old Spanish. Certainly not every aspect of the subjunctive remained the 92 same in the transition from Vulgar Latin to Old French, but the shifts that did occur began as subtle differences that can only be identified as significant to Modern French in retrospect. 3.3.1.1 The Subjunctive in Main Clauses In Old French, the subjunctive is frequently used in main clauses. Although the subjunctive verb originally only appeared ‘alone’ in the clause as it did in Latin, in Old French there is a notable rise in usage of que in front of the subjunctive verb in the independent clause. The grammatical function of this que is heavily debated: some scholars contend that the use of que invariably constitutes a subordinate clause, and thus even a seemingly independent use of the subjunctive becomes dependent on an ‘invisible’ or ‘implied’ independent clause in the speaker's mind. Other scholars disagree, and maintain that the subjunctive can appear in a main clause even if preceded by que (cf. Winters 2013, Jensen 2017). Independent uses of the subjunctive appear with and without the que in Old French, but clauses without que are more frequent. In terms of the categories of subjunctive use from Classical Latin established in this study, the optative and volitive classifications sustain strong selection preference for the subjunctive mood in Old French. These uses include wishing (true optative subjunctive), commanding (jussive and hortatory subjunctive), exhortation and advising (prohibitive subjunctive), exclamation (deliberative subjunctive), and concession (concessive subjunctive). The independent potential subjunctives-- divided by Murphy (2008) into the subcategories of expressing mere possibility, understood conditions, perceiving/thinking, and irrealis-- are more difficult to identify and analyze in Old French. Jensen (2017: 102) notes that “the condition may take the form of an independent subjunctive, mostly without que”, but does not delineate particular pragmatic contexts of use. Thus, the independent potential subjunctives may be 93 assumed to have survived through the Old French period, but it is likely that the grammaticalization of the conditional tense initiated a process of subjunctive replacement which continues into the present. 3.3.1.2 The Subjunctive in Subordinate Clauses Old French, like Old Spanish, completely recreated a new system of conjunctions based on que, and again a pattern emerges that suggests the appearance of the subjunctive in a clause is directly related to the use of a subordinating conjunction. Of the fourteen subcategories of dependent subjunctive use outlined in this paper, Old French has the highest rate of subjunctive selection for those belonging to the optative and volitive subjunctives. As was the case with Old Spanish, in Old French “volition constitutes the strongest area of the use of the subjunctive, an area where only relatively few changes have occurred down through the centuries. The norm is the subjunctive [in a subordinate clause]; the indicative occurs sporadically when the volitive element is weak” (Jensen 2017: 33). Governing verbs rise in pragmatic significance, and gradually become a major determining factor in mood selection. Verbs expressing wishing, desire, orders, pleas, exhortations, purpose, effort, prevention, prohibition, permission, concession, judgment, necessity, advice, opinion, and blame strongly prefer the subjunctive. Verbs of cognition and belief also take the subjunctive in both affirmative and negative forms. The intensity of a verb’s volitive aspect seems to directly correlate with how often it is paired with a subordinate clause in the subjunctive. Emotive contexts (including the expression of hope), however, do not select the subjunctive, again paralleling Old Spanish. Jensen (2017: 45-46) asserts: “In Old French, verbs of emotion take the indicative, focused on objective existence. [...] The subjunctive after an expression of 94 emotion is extremely rare in the Old French period.” In addition to an increased association with volition, the subjunctive in Old French also grew tightly linked with the concepts of doubt and subjectivity. Although to the modern reader an emotion seems to belong to the realm of the subjective, in Old French, an expression of emotion was akin to presenting an objective fact (i.e., the speaker reporting a factual emotional state) , and therefore required the indicative mood (cf. Jensen 2017). It is crucial to note that fear clauses does take the subjunctive, tracing back to its heritage as the optative subjunctive. Verbs expressing doubt and uncertainty also prefer the subjunctive in subordinate clauses. This category also includes verbs of certainty that have been negated. As Jensen (2017: 122) comments, “noteworthy under the category of possibility and doubt, is the seemingly automatic use of the subjunctive after a negated governing verb even when objective facts are involved”. Fear clauses, final (purpose) clauses, consecutive (result) clauses, causal clauses, concessive clauses, and comparative clauses all prefer the subjunctive, although variation occurs in accordance with differing degrees of volition, subjectivity, and/or doubt. Conditional clauses regularly select the subjunctive, but the conditional and future indicative are not uncommon. 3.3.1.3 The Subjunctive in Indirect Discourse The subjunctive is present in indirect discourse. Indirect statements expressing commands almost always take the subjunctive mood, likely due to a high level of volition. Regarding indirect questions, however, Jensen (2017: 68) reports “modal hesitation” in Old French. Volitive and dubious contexts take the subjunctive, in less straightforward scenarios the indicative and/or infinitive appear. 95 3.3.2 Modern French The Modern French subjunctive has been and continues to be a controversial topic within the field of linguistics. Two academic schools of thought have surfaced: some scholars and researchers firmly believe that the Modern French subjunctive is so reduced in form and meaning that it is essentially dead, and its lifeless morphological corpse haunts only the over-educated and overly pedantic. The second school of thought insists that the Modern French subjunctive is alive and well, and will endure with the French language itself in both form and meaning. Several theorists have taken stances in the ideological ‘middle’. In terms of morphology, as discussed in chapter 2, the Modern French subjunctive paradigm has only two productive tenses out of the original four. The following sections present the pragmatic state of the Modern French subjunctive, again relying on data from Murphy (2008). 3.3.2.1 The Subjunctive in Main Clauses The following table, adapted from Murphy (2008), is the same as Table 7 above, and presents the same subjunctive categories. Each original use of the subjunctive is color coded; the lighter the color, the stronger the subjunctive selection bias. Cells with a white background require a subjunctive almost exclusively. Cells with a light gray background take the subjunctive, but also accept the indicative. Cells with a black background have been replaced entirely by a non-subjunctive mood. Table 14: Uses of the Subjunctive in Modern French Independent Clauses Optative Subjunctive Volitive Subjunctive Potential Subjunctive Optative que les dieux l’empêchent may the gods prevent that Jussive qu’ils le disent let them tell “May” quelqu’un pourrait dire someone may say 96 Hortatory aimons notre pays let us love our country “Would” le temps manquerait time would fail Prohibitive no résistez pas do not resist “Could” on pourrait croire one could believe Deliberative qu’est-ce que je ferai what shall I do “Should/Would/Could Have” j’aurais été philosophe I would have been a philosopher Concessive j’accepte que ce soit vrait I grant that this is true Note: The subjunctive conjugations are bolded. It is clear from the table above that most of the independent uses of the subjunctive are no longer productive in Modern French. Only two types of main clause continue to exclusively take the subjunctive: the optative (for wishes and desires) and the jussive (third person commands). While in Old French the use of the complementizer que was infrequent, in Modern French the only employment of the independent subjunctive that do not require que are so called “stock phrases” like Dieu vous bénisse “[may] God bless you”, soit “so be it”, and vive la France “long live France” (Winters 2013: 353). In the volitive subjunctive category, the hortatory and concessive uses are steadily ceding to imperative and the indicative, respectively; the prohibitive subjunctive has been completely replaced by the imperative, and the deliberative by either the indicative, the infinitive, or the conditional. The most reduced category by far is the potential subjunctive, which no longer has any uses that exclusively accept the subjunctive. The “may” 97 potential for expressing mere possibilities will still take a subjunctive, but often is substituted by the indicative or conditional. The “would” (understood condition), “could” (perceiving/thinking), and “should/would/could have” (irrealis) potentials solely select the conditional mood. 3.3.2.2 The Subjunctive in Subordinate Clauses The following table, adapted from Murphy (2008), is the same as Table 8 above, and presents the same subjunctive categories. Each original use of the subjunctive is color coded; the lighter the color, the stronger the subjunctive selection bias. Cells with a white background require a subjunctive almost exclusively. Cells with a light gray background take the subjunctive, but also accept the indicative. Cells with a black background have been replaced entirely by a non-subjunctive mood. Table 15: Uses of the Subjunctive in Modern French Dependent Clauses Optative Subjunctive Volitive Subjunctive Potential Subjunctive Fear (Purpose) j’ai peur qu’il ne vienne pas I’m afraid that he won’t come Final (Purpose) je le dis pour que tout le monde le comprenne I say it so that everyone understands Consecutive (Result) il a tant ravagé la Sicile qu’on ne puisse pas la restaurer à ce qu’elle était avant he so ravaged Sicily that it cannot be restored to its former condition Adversative & Concessive quoique toute terreur me menace, je vous aiderai although all terrors hang over me, I will lend aid Characteristic il y a beaucoup des choses qui rendent plus sage there are many things which sharpen the wits 98 Causal puisque tu es mortel, ocupe-toi de ce qui est mortel since you are mortal, care for what is mortal Proviso qu’ils sé marient, pourvu qu’aucune dot ne soit donnée let them marry, provided that no dowry goes with it Circumstantial “Cum” comme le printemps approche, il faut quitter notre campement d’hiver as spring is approaching, we must move out of our winter quarters “Anticipatory” (Temporal) j’attend que la lettre arrive I shall wait until the letter comes Comparison (Conditional) je te servirai, comme si tu m’avais acheté I will serve you as though you had bought me for money Protasis of “Ideal” (Conditional) si vous disiez cela, vous vous tromperiez if you should say this, you’d be mistaken Protasis of “Unreal” (Conditional) on ne chercherait pas la sagesse, si elle ne comptait pour rien wisdom would not be sought, if it accomplished nothing Note: The subjunctive conjugations are bolded. 99 Despite the fact that Modern French employs the subjunctive much more frequently in subordinate clauses than main ones, overall usage of the subjunctive has decreased with time. The optative subjunctive continues to strongly select the subjunctive, aided by the eventual assimilation of all expressions of emotion into the fear clause structure. Proviso clauses are the second clause type that only takes the subjunctive, a phenomenon that may perhaps be explained by a relatively solid volitive connotation. Purpose clauses favor the subjunctive, especially when marked with a subordinating conjunction like pour que (a descendent of que), but increasingly select a construction with the infinitive. In the same way, adversative/concessive clauses favor the indicative unless a subordinating conjunction is used, in which case a subjunctive will follow. Clauses of characteristic and “anticipatory” temporal clauses determine mood from degree of certainty; like Modern Spanish, if the clause is certain or specific, the mood will be indicative, and if the clause is uncertain or general, the mood usually switches to subjunctive. Consecutive (result) clauses tend to be indicative due to the high level of objectivity involved in reporting a result, but the appearance of the subjunctive is not completely rare. In causal clauses, circumstantial clauses, as well as the three conditional clause types (comparison, protasis of “ideal”, and protasis of “unreal”), the subjunctive has been ousted entirely by the indicative. A probable motivation for this development in causal and circumstantial clauses is the low degree of subjectivity in these statements, as most clauses communicate information that is known or observable. The three conditional clause types now require an indicative due to a fundamental change in the formulation of conditional sentences. Begioni (2010: 13) theorizes that: “in classic French, the gradual disappearance of the Past Simple brought about the structural collapse of the Subjunctive Imperfect, which was replaced 100 by the Indicative Imperfect in the unreal subordinate”. The bond between the passé simple (the original French preterite indicative) and the imperfect subjunctive (which was originally the pluperfect subjunctive) as perfect tenses traces back to Classical Latin. The advent of the periphrastic perfect system weakened the original perfect tenses, and disrupted the subjunctive system. This disruption of the subjunctive led to the selection of the indicative in conditional subordinate clauses. 3.3.2.3 The Subjunctive in Indirect Discourse The subjunctive is no longer used in indirect discourse. Indirect statements usually take the indicative, but may also select the conditional. Indirect questions select the indicative exclusively. 3.4 Patterns of Variability and Stability From the very beginning, the story of the subjunctive has been particularly complex. Born an amalgamation of two grammatical moods, even in the time of Classical Latin the subjunctive has shown what Jensen (2017) repeatedly calls “a dual character”. Although researchers can agree that significant changes have occurred, an accurate depiction and analysis of the subjunctive has eluded linguistics for centuries. The following section will not pretend to have any definitive answers to the question of the subjunctive, but rather will attempt to unify the data and conclusions of others. First, patterns of variability and stability will be identified. Then, these patterns will be evaluated using the lens of preexisting linguistic theory. 101 3.4.1 Patterns of Stability It is considerably more difficult to pinpoint themes of stability throughout the entire odyssey of the subjunctive than it is to identify stable elements in a singular language. To find our bearings in this tumultuous sea, we must return to the site of genesis: the Proto-Indo-European subjunctive and optative. The purpose of the PIE optative mood was to articulate wishes, desires, and prayers; the purpose of the PIE subjunctive was to express hypothetical scenarios and distant future eventualities. Latin inherited both sets of pragmatic functions in its own subjunctive mood. This new subjunctive could appear in independent or dependent clauses, and had no special marker besides the verbal morphology. At this point, the evolutionary process begins, slowly but surely. An exhaustive list of the precise capacities of the subjunctive mood has already been detailed, both in this paper and in hundreds of others like it. Our focus now turns toward the singular most important factor in the subjunctive mood, perhaps the only one that has remained stable intra and interlinguistically: the binary contrast between the subjunctive and the indicative-- between ‘real’ and ‘unreal’. The essential principle is this very opposition between indicative and subjunctive. Categorizing the various uses of the subjunctive is helpful, but the true significance of the subjunctive is actually inextricably linked with the indicative; they depend on each other to define meaning, and there cannot be one without the other. At the base level, the indicative marks that which is ‘real’-- facts, truths, observations. The subjunctive, on the other hand, marks that which is ‘unreal’-- wishes, hypotheticals, subjective interpretations. This opposition in meaning is supported by an opposition in construction. That is to say, the subjunctive is also distinguished linguistically in morphology and syntax. Morphology plays an 102 important role in independent clauses, while syntax plays an important role in dependent clauses. Significant developments have occurred in the subjunctive mood, as the following section will discuss, but the fundamental idea of the indicative/subjunctive binary has remained stable. 3.4.2 Patterns of Variability The evolution of the subjunctive mood can be generalized into two types: semantic changes and syntactic changes. While these two types have been separated for the purpose of this study in reality they are so entirely connected as to be virtually inseparable. Figure 1 presents a rough timeline of shifts in the semantic facets of the subjunctive, and Figure 2 presents a rough timeline of shifts in the syntactic facets. Figure 1: Pragmatic Evolution of the Romance Subjunctive Figure 2: Syntactic Evolution of the Romance Subjunctive 103 As previously discussed, the earliest uses of the subjunctive in Latin are to convey wishes (inherited from the PIE optative) and hypotheticals (inherited from the PIE subjunctive). Classical Latin expanded these categories; wishes extended to desires, commands, concessions, prayers, and fears, while hypotheticals extended to possibilities and conditions. In the early Romance languages, more uses of the subjunctive were expanded, including judgements, opinions, doubts, and uncertainties. The modern day uses of the subjunctive have become broader still, and can be divided into the overlapping classifications of volition (wishes, desires, commands, prohibitions, concessions etc), potential (doubt, uncertainty, (im)possibilities, hypotheticals etc), and subjectivity, which combines aspects of both volition and potential (emotions, judgements, opinions, hopes, conditions etc). These categories are not intended to provide a complete list of uses, nor is this ‘timeline’ intended to represent precise chronological events. The intention is rather to illustrate the encompassing nature of the pragmatic subjunctive evolution, and to connect original uses with newer ones. In terms of the syntactic changes that have taken place, relevant events commence in Latin with the gradual upheaval of the syntactic organization of the language. In the shift from left-branching to right-branching linguistic structures, sentences that were once constructed paratactically became constructed hypothetically. This led to the increased use of subordinating conjunctions in order to mark a subordinate clause. Increased use of the subordinating conjunction system in part triggered a systemic upheaval, and the specialization (narrowing down) of subordinating conjunctions widely used. Meanwhile, the subjunctive is increasingly used with subordinate clauses. This causes a widespread association between the subjunctive mood and subordinate clauses. In turn, this causes the widespread use of the subjunctive with 104 subordinate clauses, generalizing the subjunctive in places where the indicative once appeared. As with Figure 1, this representation is not meant to provide a definitive historical analysis, but rather a rough sketch of a probable sequence of events. These semantic and syntactic developmental processes occurred simultaneously, and reinforced one another. They were also influenced by external factors. The following section reviews the cycles of reinforcement and coevolution of the meaning and formation of the subjunctive from a perspective of existing linguistic theory. 3.4.3 Theories of Language Change Every linguistic evolution may be explained by one simple formula: speakers adjust language to be more readily comprehensible. This principle is quite logical; after all, the entire purpose of language is to communicate effectively. If a language were to become so complex that it could not be easily used, communication would grind to a halt. Aitchison (2001:177-178) succinctly explains that “language minimizes opacity in that it lessens confusing ‘opaque’ situations, and maximizes transparency, in that it moves towards constructions that are clear or ‘transparent’”. Not only does this happen with sounds and morphology, as considered in the previous chapter, but also with syntax. The most obvious example of syntactic reorganization mentioned in the history of the subjunctive is the transition from left- to right-branching linguistic structures, and the heightened frequency of marked subordinate clauses. In left-branching structures and parataxis, multiple clauses are arranged without conjunctions. The use of hypotaxis and right-branching structures clarified overall sentence meaning. The same principle applies to the gradual reorganization of the conjunction system in Vulgar Latin: the diversity in subordinating conjunctions and relative pronouns, as well as complications in the Latin declension system, 105 created a linguistic environment that impeded communication. Eventually, the number of conjunctions was dramatically reduced to a set of non-declining (unchanging) forms. Even these were paired down, resulting in the widely applicable quod. In Romance, quod was replaced by que. Old French and Old Spanish each inherited que and used it to create their own systems of conjunctions, rediversifying its meaning to explicitly express different concepts for different clause types. The association between the subjunctive and subordinate clauses, as well as the subsequent expansion of subjunctive use in subordinate clauses, is a cycle that Aitchinson (2001: 178) describes as “misinterpretation”. In the process of clarifying language, “speakers tend to misanalyze a construction which has become confusing or unclear in terms of a more familiar one with superficial similarities”. The misinterpretation process is evident in the later developments of the subjunctive. After the rise in use of the subjunctive mood in subordinate clauses (due to the increase of marked subordinate clauses in general), speakers unconsciously linked the subjunctive with subordinate clauses. The use of the subjunctive mood in subordinate clauses increased as a result of speakers’ technically incorrect observation of a pattern (subordinate clause = subjunctive). In cases of subordinate clauses expressing emotion, two misinterpretations compounded, the first being subordinate clause = subjunctive and the second a misinterpretation of the subjunctive in fear clauses leading to the association of expression of emotion = subjunctive. Despite originally arising from misunderstandings, today the use of the subjunctive is dictated by patterns associating it with a subordinating conjunction or with governing verbs expressing ‘unreality’ like doubt or emotion. The following chapter investigates the current state 106 of the subjunctive in Modern Spanish and Modern French, particularly concerning second language and heritage learners, and anticipates future developments of the subjunctive mood in Romance. 107 “‘Speak in French when you can’t think of the English for a thing-- turn out your toes as you walk-- and remember who you are!’”-The Red Queen, Through the Looking Glass by Lewis Carroll 4. THE SUBJUNCTIVE IN MODERN ROMANCE AND SECOND-LANGUAGE LEARNERS Having concluded the chronological review of the forms and functions of the Romance subjunctive, the next leg of our journey is to stop and smell the roses-- to explore the present context of the subjunctive in Modern Spanish and Modern French. This chapter pays particular attention to studies conducted on subjunctive selection that divide the entire language speaking population into three groups: native speakers, or adults who have been speaking the target language from birth; first language learners, or children in the process of learning their first language (L1 learners); second and third language learners (L2 and L3, respectively), or people who are learning the target language as their second or third language; and finally heritage learners (HL), a term which usually refers to children or young adults who live in a bilingual household, and grow up learning and speaking two languages. The insights of exploring the use of the subjunctive within these groups are relevant for different reasons. Native speakers represent the current “standard” in a language, and provide a basis for comparison of use. Children learning their native language can illuminate the complex processes involved in acquiring a language, because they learn parts of the language sequentially. Second and third language learners, like children, acquire a new language in phases, but they can also provide valuable insight on how multiple languages interact with one another throughout the 108 language acquisition process. Heritage learners are a diverse population, but generally they acquire their first language in school, and their second language (the target language studied) informally at home. Heritage learners can thus help researchers understand the mechanics of language acquisition that occur without explicit instruction. Many studies, as well as this chapter, draw on the experiences of all four groups of people in order to inform conclusions on the nature of language acquisition. The following sections will examine the current state of affairs in Modern Spanish and Modern French, including how the subjunctive mood is used and interpreted in both languages, and then investigate research on subjunctive acquisition and usage of L2s, L3s, and HLs. Data from these studies is critical to achieve more accurate comprehension of the subjunctive. Every language learner acquires a language in steps and makes errors as they acquire a language; studying the order of acquisition and the specific mistakes of learners can reveal patterns in language use, and even the extent to which an additional language can interfere or support the learning process. Research results also provide information that can be used to identify and analyze the “strong” and “weak” spots in a language, and predict future evolutions of current trends. 4.1 Spanish “For learners of Spanish the acquisition of the subjunctive forms and their meaning continues to be one of the benchmarks of success”, begins Collentine (2010: 39). This so-called “success” is based on comparison between Spanish language learners and native Spanish speakers, wherein the ultimate goal of learners is to ‘speak like a native’. In terms of the subjunctive, this goal is a 109 lofty one, one that likely will take years of study and practice to achieve. But how do modern native Spanish speakers actually use the subjunctive? This section first reviews the usage of the subjunctive in Modern Spanish, as well as several interpretations of the subjunctive mood. Then, these findings are compared with those on the complex relationship between Spanish language learners and the subjunctive. 4.1.1 Modern Spanish Interpretations of the Subjunctive As detailed in the previous two chapters, the Modern Spanish subjunctive has changed significantly from its origins in Classical Latin, both in morphology and syntax. The subjunctive system, however, continues to function productively, with a relatively high level of complexity. This complexity has not gone unnoticed. Several studies have reported the neutralization of the subjunctive by native Spanish speakers-- that is, the substitution of the subjunctive mood for the indicative. In a comprehensive review of four studies, Faingold (2003:78) states that “[native] Spanish speakers hypothesize, express desires and wishes, and so on, in the indicative instead of the subjunctive”, and elaborates that: Latin American and Iberian Spanish speakers show variation and substitute the less-marked future, present, and present-perfect indicative for the more marked present and present-perfect subjunctive; they also substitute the less-marked past indicative for the more-marked past subjunctive. In fact, in an analysis of the Corpus del español, “a corpus of more than 21,000,000 words comprising native-speaker samples of both written and spoken Spanish from a variety of registers”, Collentine (2010:39) reports that: The data suggest that, whether in oral or written language, the proportion of subjunctive forms native speakers produce is small compared to other paradigms/conjugations, such as the present indicative, imperfect, or preterit. This analysis shows that the subjunctive, whether in the present or the imperfect, comprises only about 7.2% of all verb forms. 110 With these results in mind, it might seem like a logical assumption to predict that the subjunctive mood is coming to the end of its lifespan in Spanish. Researchers, however, disagree, and assert that the subjunctive is not at risk of extinction. Not only does the Spanish subjunctive hold a position of societal prestige, and is associated with more educated speakers (cf. Faingold 2003, Collentine 2010), but it is also argued to form a critical part of the communicative capabilities of the Spanish language as a whole. Ahern et al. (2017: 1) contend that “the rich morphology of Spanish, such as that of tense and verbal mood, encodes a range of features leading to diverse contextual effects on interpretation…”. The particular nuances of the Spanish subjunctive, while challenging to articulate and categorize, can reveal the speaker’s attitude, viewpoint, and even conception of reality (cf. Vesterinen 2013). The binary distinction between ‘real’ and ‘unreal’ inherent in the opposition between the subjunctive and the indicative provides an essential psychological interface for expression. Gregory (2001: 100) explains that: The insight that mood is the connection between language and communication is a crucial one. Communication involves cognitive processing and interactive, interpersonal, contextual and social influences. Language and cognition are inextricably linked. We view sentence structure, as well as discourse structure and coherence, as a function of natural cognitive processes. The grammatical category of mood can be considered the point of intersection of the cognitive and the interpersonal in communication and is, therefore, a high-level strategy for interpreting reality and, subsequently, making it linguistically manifest. [...] It is not just the subjunctive that carries meaning, but the fact that there is a mood contrast. It is not possible to define subjunctive by itself, for it is an inextricable element of the mood system. Therefore, despite the relative infrequency of the subjunctive overall, it remains a vital linguistic tool. Native speakers regularly interact with the subjunctive system as they regularly interact with the Spanish language. Even if an individual avoids personal use of the subjunctive, the selection or preference of the indicative still engages the subjunctive, since, as discussed in the previous chapter, the existence and significance of the indicative depends on that of the 111 subjunctive and vice versa. Given the prevalence and relevance of the Spanish subjunctive, it comes as no surprise that acquisition of the subjunctive is used as a “a “benchmark of success”. 4.1.2 Language Learners and the Subjunctive Due to the multiplicities present in the subjunctive system, it also comes as no surprise that many Spanish language learners encounter difficulties with the topic. Even academic papers on the topic have a tendency to introduce the subjunctive as a notoriously confusing subject, as does Glick (2021: 151), who begins: “el uso del subjuntivo es uno de los temas que más problemas causa a los estudiantes de español como lengua extranjera” (“the use of the subjunctive is one of the topics that causes the most problems for students studying Spanish as a foreign language”). This observation is supported by several studies on subjunctive acquisition. Faingold (2003: 75) reports that language learners display high rates of subjunctive neutralization: English speakers learning Spanish exhibit [...] the less-marked infinitive and, more often , the present indicative, as well as other structures, in place of the more-marked present subjunctive in 79.5 per cent (first group) and 60 per cent (second group) of the cases where the present subjunctive is obligatory, and [...] they show the less-marked present subjunctive, the infinitive, the present indicative, the future, and, more often, the past indicative, as well as other structures, for the more-marked past subjunctive in 94 percent (first group) and 89 percent (second group) of the cases where the more-marked past subjunctive is obligatory. Heritage speakers specifically also encounter similar obstacles in subjunctive mood selection, notes Giancaspro (2017: ii): “Unlike Spanish-dominant speakers, whose knowledge of subjunctive mood appears to be largely categorical, HSs’ [Heritage speakers’] subjunctive mood knowledge is characterized by great variability, both in production and comprehension”. The dramatic difference between native speakers and language learners can be explained in part by first-language interference. While native speakers start learning a particular language “from scratch”, with no basis for comparison, in many cases language learners have already 112 completely acquired a first language before starting to study another. Knowledge of the first language has a tendency to “interfere” with the learning of the second language-- learners expect features of the second language to mimic those of the first, and thus can misinterpret the second language in structure and meaning. This first language knowledge, on the other hand, can also positively interfere with the second language-- that is, if an element in the first and second languages are similar, language learners will have an easier time acquiring it. Sanchez-Naranjo (2010: iii) summarizes this concept: Results indicate that subjunctive adjuncts present difficulties in L2 acquisition even for advanced L2 learners. Although they exhibit sensitivity to certain subjunctive features and contextual meanings, data reveal that convergence and non-convergence were primarily determined by L1 influence on L2. Crucially, those features absent from the L1 give rise to greater efforts and difficulties in L2 form-meaning mappings of mood selection. Interestingly, this positive interference also occurs when the second and third language share the similarity, even if the first language does not, as was the case in Elordi (2012: 115), where English speakers learning Spanish and French evaluated on their competence in the subjunctive mood: “positive crosslinguistic interference occurs both from the direction of the L2 to the L3 and from the L3 to the L2 […] Such results also suggest that the adult L2 learners are better able to acquire an interface phenomenon when they are also learning an L3 which uses it in the same way…”. Positive or negative interference aside, studies also demonstrate that, contrary to popular belief, the subjunctive “benchmark for achievement” of native-like proficiency is not out of reach for language learners. Gudmestad (2013: 1) concludes that: The results show that native speakers use the present, imperfect, and pluperfect subjunctive forms more frequently than other subjunctive forms and that they use these three forms in a range of linguistic contexts. The analysis also demonstrates that learners use these three forms in largely the same contexts as the native speakers and that they 113 restructure the strength of their form-meaning associations as they become more proficient in the target language. Elordi (2012: 115) agrees, stating that “our findings lead us to join the side of the debate that believes a near-native competence is possible despite the difficulties acquisition of these phenomena entail”. Heritage learners can also successfully acquire the subjunctive, notes Viner (2018: 1): “we [...] challenge widely-accepted notions surrounding heritage languages, namely simplification, attrition and incomplete acquisition, concluding that all three lack the theoretical means to accurately describe the complex bilingual situation in NYC”. 4.2 French Through the centuries, the French subjunctive has acquired a notorious reputation of complexity, and has therefore become a sort of linguistic Pandora’s box for researchers. This characterization continues in the research of today. Comeau (2020: 21) introduces the topic by stating that “there is arguably no other grammatical feature of French which has spilled more ink than the subjunctive mood”, while Poplack et al. (2013: 140) calls the subjunctive “the most elusive and mysterious feature of French grammar”. As discussed in the previous two chapters, the French subjunctive has undergone extensive developments since its inception. How is the Modern French subjunctive used by speakers today? The following sections examine modern mood selection and interpretation by native speakers and language learners. 4.2.1 Modern French Interpretations of the Subjunctive The substantial quantity of research on the French subjunctive, however, appears disproportionate to the statistics on its actual employment in today’s language. McManus & Mitchell (2015: 45) report that: 114 In a corpus study based on 53,265 clauses of spoken and written native-speaker French from a variety of different genres (e.g. conversations, interviews, magazines, novels), O’Connor DiVito (1997) found that only 2% (n=342) of 16,236 spoken clauses and 3% (n=1034) of 37,029 written clauses contained a Subjunctive form. These percentages fell to 1% (n=233) and 2% (n=789), respectively, when only clearly identifiable Subjunctive forms were counted (e.g. fasse, finisse). These are substantially lower than percentages reported for Spanish… Besides low overall rates of usage, the French subjunctive faces an additional complication: in spoken and written language, some verbs no longer distinguish morphologically between the indicative and the subjunctive. One-third of all subjunctive forms are identical in their indicative counterparts; for -er conjugation class verbs, the present subjunctive and the present indicative only have separate forms for the first and second person plural (McManus & Mitchell 2015: 43). McManus & Mitchell (2015: 46) also present that “the verbs être (‘be’), avoir (‘have’), pouvoir (‘be able to’), faire (‘do’) and regular –er verbs (e.g. regarder ‘watch’) account for 78% of all Subjunctive forms in the corpus”, and that 53% of all subjunctive forms are comprised of these four verbs (être, avoir, pouvoir, and faire) alone. Although the data may indicate that the days of the French subjunctive are numbered, grammarians and linguistic institutions continue to assert its relevance and uphold the use of the subjunctive as an academic and social standard. This position is supported by research like Poplack et al. (2013: 156), which notes the prestige of the subjunctive compared to the indicative, and comes to a bold conclusion: only the attempts at categorization of subjunctive use have varied, not the actual subjunctive mood selection. They contend that: Beyond the highly frequent and frozen prefabs containing subjunctive morphology, speakers have stored subjunctive schemas which they fill with material which is determined not by considerations of meaning, but by linguistic, stylistic and extra-linguistic community norms. It will by now be obvious that the grammar of the subjunctive referred to in the title [“The evolving grammar of the French subjunctive”] as evolving is prescriptive grammar. The grammar of actual usage, in contrast, has remained remarkably stable, despite pervasive variability. (191) 115 This defense of the French subjunctive has been echoed by other researchers and theorists, to the point that a certain quote from Brunot (1936: 519) has become a sort of literary cliché in the genre: "Rien, absolument rien, ne fait prévoir que la forme du subjonctif soit menacée de périr" (“Nothing, absolutely nothing, predicts that the form of the subjunctive is threatened with perishing”). In light of this perspective, it is no wonder that modern French grammars continue to devote pages upon pages on explanations of the subjunctive and its use. 4.2.2 L2 Learners and the Subjunctive Compared to the rich corpus of studies investigating the relationship between language learners and the subjunctive in Spanish, the current research on the French subjunctive is not as all-encompassing in terms of scope or depth. Up-to-date research on L2 learners is hard to find, and research on heritage speakers is almost non-existent. From the information available, however, one conclusion is clear: as in Spanish, French language learners tend to struggle with the acquisition of the subjunctive mood. In a longitudinal study on French learners, McManus & Mitchell (2015: 183) find that the subjunctive is a “late-emerging” grammatical feature, and that the neutralization of the subjunctive is common even in advanced learners: Although our learners, and in particular our study abroad learners, have gained considerable fluency, are very communicative, and have developed an interlanguage that is highly grammaticalised on other morpho-syntactic forms, the subjunctive still alludes [sic] them, and this in spite of many years of learning French and its common usage in the variety of French to which they were exposed, namely metropolitan French. Although they produce a range of syntactic contexts where the subjunctive could be expected, the form still remains to be acquired, such that, in relation to subjunctive-inducing contexts at least, the learners’ syntactic development is more advanced than their morphological development. Furthermore, in no way does the acquisition of form precede function in the case of the subjunctive, such that we find no exemplars of its usage in contexts where it is not required, and only minimal usage in contexts where it is required. 116 This phenomenon may be explained in part by findings of a second study, Dudley & Slabakova (2021), which evaluated language learners’ knowledge of the obligatory French subjunctive in real-time contexts. Dudley & Slabakova (2021: 1) affirm that during real-time processing, L2 learners experience challenges applying grammatical knowledge, that is, translating information that they have learned into actual linguistic production: Data from an acceptability judgment task and an eye-tracking during reading experiment revealed that L2 learners had knowledge of the subjunctive, but were unable to apply this knowledge when reading for comprehension. Such findings therefore suggest that L2 knowledge of the subjunctive, at least at the proficiency levels tested in this study, is largely metalinguistic (explicit) in nature and that reduced lexical access and/or limited computational resources (e.g., working memory) prevented learners from fully utilising their grammatical representations during real-time processing. Despite the presence of evidence that may be interpreted ‘pessimistically’, research also suggests that when language learners eventually become comfortable with the subjunctive, they can reach the same “benchmark of success” of native-like competency. Elordi (2012: 78) concludes that “worthy of note is the fact that there are no significant differences between the way that NSs [native speakers] interpret mood and the way that NNSs [non-native speakers] interpret mood”. 4.3 Social Factors in Theories of Language Change The previous two chapters have presented the argument that because all language is based on a finite set of patterns, all language change consequently involves a process of a breakdown in the organization of a certain pattern and attempts to reorganize said pattern. In other words, languages strive to achieve linguistic balance; when a language’s system of patterns becomes unbalanced, change will occur to rebalance the system. It has also been contended that linguistic evolution is not random, but rather depends on the existence of “weak spots”, or points of 117 vulnerability, which are then exploited. As introduced above, in the work of Aitchison (2001), two major categories of change are outlined: internal psycholinguistic factors, which determine the “weak spots” in a language; and external social factors, which accelerate the exploitation of “weak spots”. This section will illustrate several social factors that influence language change, including the relationship between language learners and linguistic evolution. It has often been theorized that contact between languages has a significant effect on how those languages develop. Geographic areas where several different languages, for example, have been the subject of much research, leading to the observation that “inhabitants of such regions are frequently bilingual or have a working knowledge of the other language(s) in the area, in addition to their native language. In this situation, the languages tend to influence one another in various ways. The longer the contact, the deeper the influence” (Aitchison 2001: 138). This concept is proven by even the surface history of the Romance languages; although the original ‘standard’ language imported into each of the Roman territories was Latin, over time, regional differences accumulated until the variety of Latin being used no longer resembled the Classical standard, and a new language emerged. The territories occupied by the Romans were also inhabited by other cultures, like the Muslims in Spain and the Gauls in France. Although Roman colonists lived in the territories, the vast majority of denizens belonged to other ethnic groups, and learned Latin as an additional language. The unique ‘character’ of each of the modern day Romance languages can be traced back to the individual sociocultural and linguistic contexts of the ancient Roman territories. Aitchison (2001: 137) elaborates on this principle, which is also called substratum theory: “when immigrants come to a new area, or when an indigenous population learns the language of newly arrived conquerors, they learn their adopted language 118 imperfectly. They hand on these slight imperfections to their children and to other people in their social circle, and eventually alter the language”. Interestingly, sometimes it is not ‘mistakes’ that are passed on, but overcorrections-- when speakers, usually with a multilingual or immigrant background, change their language use in an attempt distance themselves from ‘uneducated’ speakers. This is especially prominent in the realm of pronunciation, where children descended from immigrants exaggerate and over-enunciate words to avoid speaking with a ‘foreign’ accent. Although no study has specifically investigated the role of contact-induced language change in the historical development of the Romance subjunctive, substratum theory can easily be connected to research in the modern field of language acquisition. Aitchison (2001: 141) notes that the so-called “imperfections” that language learners pass on are deeper than vocabularic errors: “when people learn a new language, they unintentionally impose some of their old sound patterns, and to a lesser extent, syntax”. This, in other words, is the principle of language interference, and first language interference in particular. The subjunctive, a syntactical element of the Romance languages, is most difficult conceptually for language learners when their first language, the language of unconscious comparison, does not have a grammatical equivalent. As numerous studies on French and Spanish have shown, English language natives (who do not have a readily identifiable subjunctive morpheme) construct sentences in the target language as they would in their first language, that is, using the indicative or conditional where a subjunctive would traditionally be obligated. Substratum theory and language interference (especially from English) have the potential to considerably influence the future of the Romance subjunctive. 119 Another social theory reported by Aitchison (2001: 147) is the functional view of language change, or the precept that “language alters as the needs of its users alter”. While generally applied to the assimilation of ‘foerign’ vocabulary into a language (for example, the French les jeans and le sandwich and the Spanish los jeans and el sandwich taken from English), it is not uncommon for more complex linguistic features like grammar and syntax to be adapted. The functional view of language change can be supported by evidence from the development of the subordinating conjunction system in Vulgar Latin and early Romance. Classical Latin employed an extensive array of subordinating conjunctions, which conjugated in accordance with context and in agreement with the gender, number, and case of the element being described by the subordinate clause. The use of a given subordinating conjunction in a certain form provided necessary information, especially considering the fact that Classical Latin used left-branching structures and parataxis, meaning that often the “subordinating clause” would come before the main clause in a sentence. In English, a typical sentence directly translated from paratactic, left-branching Classical Latin would look roughly like this :“whom I love which is on the shelf the girl the picture painted”. English, a hypotactic and right-branching language, would prefer the sentence “the girl whom I love painted the picture which is on the shelf”. The Classical Latin sentence could only be pieced together by matching the subject of the clause and the subordinating conjunction, so the comprehensive system of subordinating conjunctions was needed. In English, most context is assumed from the formula ‘main clause, subordinate clause’, so a conjunction system like that of Classical Latin is not needed. The needs of speakers changed from Classical Latin to Vulgar Latin. Among other factors (cf. Murphy 2008 for an in depth discussion), the transition from parataxis to hypotaxis 120 eliminated the social need for the Classical Latin conjunction system. As previously discussed, the most commonly used conjunctions first became generalized in use (did not need to be conjugated to match the subject, and instead one form could begin any clause). The formerly long list of forms became incredibly short, and shorter still with the specialization of quod as the ‘end-all, be-all’ subordinating conjunction. This suited the needs of the language users at the time. However, in early Romance, the needs of language users changed. More specific conjunctions were needed to express different contexts (e.g. “in order to”, “after”, “by”), a phenomenon that could perhaps be connected to the increase in use of the subjunctive in subordinate clauses. Thus language changed with the needs of language users: French and Spanish each individually developed their own new system of conjunctions, expanding on the original quod. Specifically in the history of the subjunctive in Romance, the functional view of language change is supported in the changes in subjunctive morphology. For example, when the original Classical Latin perfect system fell into disuse, there was still a social need to express contexts that linguistically required the perfect tenses. An old but relatively infrequent construction was adapted into the first periphrastic perfect construction in the indicative, from which it spread into the rest of the indicative and the subjunctive, effectively filling a social need. The invention of the future and future perfect subjunctive in Old Spanish also corroborates the functional view of language change: speakers required a way to express eventualities in the future within the realm of the subjunctive, so morphology was recycled to construct the future and future perfect. With 121 time, however, this function was not needed, as speakers began to substitute the present subjunctive for the future. It is important to reiterate that these external factors of language change, both language contact and social need, do not expressly cause linguistic evolution. First language interference or a general want of linguistic features appear regularly, but in the majority of cases do not lead to widespread linguistic development. Only those external factors which coincide with places of vulnerability in a language will induce change. For instance, consider the Latin subjunctive perfect system. Latin already had a linguistic “weak spot”-- certain vowels and consonants were vulnerable to change in pronunciation, or syncopation altogether. When the basic sounds of Latin changed to make for easier and more readily accessible pronunciation, the “weak spot” became larger, because the tenses of the subjunctive became indistinguishable from those of the indicative. Social need promoted the use of the periphrastic perfects, which has the same meaning as the synthetic perfects but could not be confused auditorily. Eventually, the subjunctive was reorganized permanently with solely two-word perfect forms. Without the original vulnerability in pronunciation, the perfect system would not have come into use-- it would not have been able to exploit a weak spot. Aitchison (2001: 147) summarizes accordingly: “social need has accelerated a tendency which has been in existence for a considerable number of years. It did not in itself instigate a change, but is merely carrying an ongoing one along a little faster”. The following chapter will take these concepts, as well as those discussed in chapters two and three, into account in an attempt to predict the future evolutions of the subjunctive in Modern Spanish and Modern French. 122 “‘It seems very pretty,’ she said when she had finished it, ‘but it is RATHER hard to understand!’”- Alice, Through the Looking Glass by Lewis Carroll 5. THE FUTURE OF THE SUBJUNCTIVE Just as it is easy for one to sit at the far edge of history and judge the past in retrospect, so is it easy to stand at the beginning of a new era and make predictions about the story yet to come. As with the weather, forecasts of language change are often reported and rarely vindicated. Also like the weather, however, it is possible for one to foresee a storm by looking up at the clouds. Predicting change is not an exact science, but a science nonetheless; the more distant a prognostic, the less likely it will come true, but with study and experience, patterns of change become more visible and accurately identified. This last chapter attempts to utilize the evidence and analysis presented in each previous chapter to anticipate future developments of the subjunctive. The potential linguistic evolution of the Spanish system is evaluated first, followed by that of French. Finally, conclusions on themes from the entire paper are presented. 5.1 The Future of the Spanish Subjunctive Spanish, a “conservative” language, appears inclined to remain relatively close to the Classical Latin subjunctive morphologically and syntactically. Each of the four original tense distinctions-- the present, the imperfect, the perfect, and the pluperfect-- retain distinctive morphology in all persons and are actively employed in Modern Spanish. The subjunctive is used both in main and subordinate clauses, although in main clauses it is usually preceded by the subordinating conjunction que. 123 Pragmatically, the subjunctive continues to be a productive paradigm. While the analysis in Chapter 3 indicated the “loss” of several functions of the subjunctive in independent and dependent clauses, many elements still strongly prefer the subjunctive, especially within the realm of the ‘unreal’. True optative (wishing) and volitive (commanding) main clause capacities have stayed a hallmark of the subjunctive since the times of Classical Latin, and do not appear to be in linguistic competition with any other forms. The ‘unreal’ extends into conditional clauses describing hypotheticals, where the protasis (the first section of the sentence) of both “ideal” and “unreal” conditions, as well as comparison conditional clauses only accept subjunctive forms. The overlapping contexts of the association of the subjunctive with subordinate clauses, as well as the gradual conflation of fear clauses (an optative use of the subjunctive) with clauses generally expressing emotion have led to an extension of subjunctive usage in subordinate clauses. The psychosocial connection between the unreal with the subjunctive mood has been firmly established in the overall ‘canon’ of the Spanish language for centuries, and it is unlikely that this will change in the near future. The subjunctive doctrine is also meticulously defended by social and political establishments of language, like the Real Academia Española (the Royal Spanish Academy), which was founded in 1713 as, in its own words, “una institución cultural dedicada a la regularización lingüística entre el mundo hispanohablante” [“a cultural institution dedicated to linguistic regularization throughout the Spanish-speaking world”]. Worldwide education, in addition to a linguistic authority as a source of language standardization, decreases the efficacy of linguistic “experimentation”, which in turn decreases the odds of “innovation”, unless associated with social prestige. These circumstances are parallel to those of the Roman Empire. 124 The language of colonization spread through an expansive and incredibly culturally diverse landscape, but no significant regional variances emerged while the “standard” language continued to be widely enforced in writing and formal speech. In colloquial spoken language, however, regionalisms and socioeconomic class had a greater influence on linguistic variation, with speakers in the lower classes and furthest from Rome geographically speaking a Vulgar Latin that already differed from the Classical standard. These beginnings of linguistic fragmentation between Classical and Vulgar Latin, as well as the fragmentation between regional varieties of Latin, only intensified over the centuries. When the Roman Empire fell, the former territories were left to their own linguistic devices, and thus germinated the seeds that would blossom into entirely distinct languages. The Spanish-speaking world of today is the linguistic footprint of Spain’s colonial empire. Territories across the globe retain Spanish as an official language, with ‘Spain Spanish’ still functioning as the “standard”. Language varies regionally between and within countries, of course, but most (if not all) of that variation consists of vocabulary and surface-level syntax. As previously discussed, morphologically the subjunctive continues to be a productive paradigm in every major Spanish-speaking country. Although the subjunctive is not taught explicitly to native speakers, it is still acquired and employed frequently even in spoken language. One possible explanation is that the subjunctive morphology and syntactic constructions have a high level of prestige, and are associated with education and socioeconomic status. Spanish language learners are taught the subjunctive explicitly, and emphasized in textbooks as a key component of Spanish fluency. This continued importance of the subjunctive belies continued usage into the future. 125 5.2 The Future of the French Subjunctive Despite extensive controversy that has surrounded and will continue to surround the French subjunctive mood, this study asserts that it is not in imminent danger. It must also be acknowledged, however, that there is a logical basis for the condemnation of the subjunctive mood. Morphologically and syntactically, the position of French as an “innovative” language is clear; while all four of the original tense distinctions from Classical Latin still exist, the imperfect and pluperfect subjunctives are confined to literary uses, and are rarely (if ever) employed in everyday language. The present and perfect (today known simply as the “past” subjunctive) lack a singular morphological marker to distinguish conjugations from the indicative. As previously discussed, one-third of these subjunctive forms are identical from those of the indicative. When the subjunctive does appear, there is a high chance that it is one of only four verbs (être, avoir, pouvoir, or faire). In terms of pragmatic meaning, French has also experienced reductive developments. In main clauses, the conditional has usurped most functions that previously only selected the subjunctive. In subordinate clauses, the indicative has largely replaced the subjunctive. On the other hand, several factors promote the survival of the French subjunctive. True optative and volitive independent clauses continue to strongly prefer the subjunctive, as do subordinate clauses expressing emotion. There persists a relatively strong connection between the subjunctive mood and the subjective, as in Spanish. Furthermore, like Spanish, French boasts the prestigious Académie française, founded in 1635 with the express purpose of “travailler, avec tout le soin et toute la diligence possibles, à donner des règles certaines à notre langue et à la rendre pure, éloquente et capable de traiter les arts et les sciences” [“working, with all possible 126 care and diligence, to give certain rules to our language and to make it pure, eloquent, and capable of handling the arts and sciences”]. The Académie française continues to act as the highest authority on the French language, and definitively establishes norms and standards of the French language. The institution of the French language is also maintained through the global education system, which strongly discourages all linguistic divergences. As noted in Poplack et al. (2013), the subjunctive mood is associated with education and prestige. This position of linguistic power, as well as the considerable influence of sociopolitical linguistic institutions, must be taken into account when considering the future of the French subjunctive. 5.3 Conclusions: The Future of the Romance Subjunctive The most compelling argument in favor of the continuation of the Romance subjunctive is rooted in the functional view of language change-- the theory that, as Aitchison (2001: 147) reiterates: “language alters as the needs of its users alter”. In other words, language changes when and if the needs of users change. In the case of Romance languages, the conclusion is simple: regarding the subjunctive mood, the needs of speakers have not changed. The subjunctive mood serves a definite and necessary linguistic purpose: to express the ‘unreal’. The domain of what is considered ‘unreal’ has evolved and will continue to do so (consider the expansion to include all emotive expressions), but in the Romance family the subjunctive mood will likewise continue to express that ‘unreal’. The volitive, the subjective, the potential: these concepts did not come into existence after the subjunctive mood had already been formed. Rather, the subjunctive mood was formed to aid in the communication of these complex eventualities. Humans will continue to have a need to discuss hypothetical situations, their wants, their wishes and hopes and dreams 127 and fears. In a way, the subjunctive represents what makes us human. “There is not a preferred natural direction of language change,” concludes Hamans (2017: 1). “Morphological change is not a result of mechanical, predictable processes, but of the behavior of language users.” In other languages, English among them, the subjunctive does not exist morphologically, or is in a state of advanced decline. Yet, as humans, there still persists a need for the linguistic expression of the ‘unreal’. The process of the gradual disappearance of the subjunctive that has occurred in these languages, also called desemantization, is deeply intertwined with the linguistic theory of economy. Studies like Poplack et al. (2018) posit that over time, the verbal forms of the subjunctive mood, originally the only ‘unreal’ modal marker in a language, became less independently significant, and increasingly associated with the syntactic structures in which the verbs occurred, with the use of verbs in the subjunctive eventually becoming restricted to certain formulaic expressions. Pragmatically, the meaning of the ‘unreal’ in these structures was reinterpreted, and the expression of modality was carried not only by the subjunctive verbs, but by other linguistic elements (for example, the helping verbs “could” and “would” in English). But in language, having multiple elements signifying the same concept creates redundancy and complicates communication. In accordance with linguistic economy theory, one of the two elements becomes less productive and is slowly eliminated from language. Thus, while the ‘unreal’ itself can still be articulated in languages like English, the subjunctive morphology itself is no longer the linguistic vehicle of conveyance. Poplack et al. (2018), as well as other Romance scholars, have applied this so-called “cline of desemantization” to the Modern Romance languages, wherein Spanish is placed comparatively closer to Latin, while French appears much further along the “cline”. The logical 128 assumption, then, leads to the conclusion that the French subjunctive is threatened with extinction, with Spanish sure to follow. Evidence presented earlier in this study could support this conclusion. I, however, will argue the opposite, concurring and expanding upon the work of Murphy (2008). The developments seen in the Romance subjunctive system have been extensive, but do not constitute the complete desemantization of the subjunctive morphology. Murphy (2008: 289) determines that an academic reassessment of the linguistic dogmas surrounding the subjunctive is necessary: Instead of asking Why is the subjunctive less frequent in Romance languages than in Latin?, we should be asking precisely the opposite: Why was the subjunctive so pervasive in Latin? The underlying assumption behind the first of these questions reveals the traditionally-held view that Latin was an ideal, highly-motivated system, with all subsequent changes representing language decay. My analysis has revealed, however, that subjunctive usage in Latin had become not only inconsistent, but also somewhat conditioned by its syntactic environment and was, therefore, lacking the semantic motivation linguists generally ascribe to the paradigm. This “inconsistency” in Latin, asserts Murphy (2008), is evidenced in the range of attempts over the centuries to successfully encompass the subjunctive system with a set of rules governing its pragmatic and syntactic usage. With time, two major components of the subjunctive emerged in Latin: the independent subjunctive in the main clause, and the dependent subjunctive in the subordinate clause. Each clause type had various pragmatic functions ascribed to the subjunctive, as illustrated in the previous chapters. For this reason, linguists have generally assumed that the subjunctive was quite productive in Latin. Murphy (2008), however, turns this assumption on its head, and theorizes that this frequency of use demonstrates a loss in pragmatic potency. In late Vulgar Latin and early Romance, the subjunctive paradigm became semantically “remotivated” as speakers reassessed the expression of modality. While certain uses of the subjunctive do not 129 continue from Classical Latin into Modern Romance, other uses, especially those involving emotion, were expanded upon. As we reach the end of the odyssey of the subjunctive, we must consider the small and seemingly unrelated threads of history that have been intricately woven together to create a tapestry that spans centuries and generations of speakers in three different languages. The morphological inheritance of Modern Spanish and Modern French is a clear homage to the comprehensive subjunctive system of Classical Latin, itself a reinterpretation of the Proto-Indo-European optative and subjunctive moods. The syntaxes of Spanish and French have changed extensively, but remain connected to that of Latin. Pragmatically, the meaning of the subjunctive appears unstable, but as proposes Poplack et al. (2013), these variations are the attempts of scholars to find logic in an inherently illogical concept. The core of the subjunctive has remained consistent throughout the tapestry, especially the winding, unbroken threads of the true optative and volitive. The subjunctive draws meaning from its binary contrast with the indicative; one cannot exist without the other. In order for something to be ‘real’, something else must be ‘unreal’. In order for there to be a world of the indicative, there must be a world of the subjunctive “through the looking-glass”. 130 BIBLIOGRAPHY Alba de la Fuente, A., Enríquez, M. C., & Lacroix, H. (2018). Mood Selection in Relative Clauses by French–Spanish Bilinguals: Contrasts and Similarities between L2 and Heritage Speakers. Languages, 3(3), 31. https://doi.org/10.3390/languages3030031 Aitchison, J. (2001). Language change : progress or decay? (Third edition.). Cambridge University Press. Alkire, T., & Rosen, C. G. (2010). Romance languages : a historical introduction. Cambridge University Press. Batstone, William. “Latin Grammar.” Department of Classics, The Ohio State University, 2022, classics.osu.edu/Undergraduate-Studies/Latin-Program/Grammar. Accessed 20 Apr. 2022. Bauer, Brigitte, and Jonathan Slocum. “Old French Online: Series Introduction.” Linguistics Research Center, The University of Texas at Austin, 2022, lrc.la.utexas.edu/eieol/ofrol. Begioni, L. (2010). L’EVOLUTION DU SYSTÈME DE L’IRRÉEL DU LATIN CLASSIQUE AUX LANGUES ROMANES. CONVERGENCES DÉCALÉES DU FRANÇAIS ET DE L’ITALIEN. (French). Studii de Ştiintă Şi Cultură, 6(2), 13–22. Becker, M. G., & Remberger, E. (Eds.). (2010). Modality and mood in romance : Modal interpretation, mood selection, and mood alternation. ProQuest Ebook Central https://ebookcentral-proquest-com.proxy.mtholyoke.edu:2443 Bossé Andrieu, J. (1983). Le subjonctif. In Abrégé des règles de grammaire et d’orthographe (ARGO). Presses de l’Université du Québec. Collentine, J. (2010). The Acquisition and Teaching of the Spanish Subjunctive: An Update on Current Findings. Hispania, 93(1), 39-51. Retrieved August 3, 2021, from http://www.jstor.org/stable/25703392 Comeau, P. (2020). When a linguistic variable doesn’t vary (much): The subjunctive mood in a https://doi.org/10.3390/languages3030031 https://ebookcentral-proquest-com.proxy.mtholyoke.edu:2443/ http://www.jstor.org/stable/25703392 131 conservative variety of Acadian French and its relevance to the actuation problem. Journal of French Language Studies, 30(1), 21-46. doi:10.1017/S0959269519000255 Connors, K. (1974). The Subjunctive in Decline: The Case of French. Dreer, Igor. Expressing the Same by the Different : The subjunctive vs the indicative in French, John Benjamins Publishing Company, 2007. ProQuest Ebook Central, https://ebookcentral.proquest.com/lib/mtholyoke/detail.action?docID=622555. De Mulder, W., & Lamiroy, B. (2012). Gradualness of grammaticalization in Romance. The position of French, Spanish and Italian. Grammaticalization and language change: New reflections, 199-226. Faingold, E. D. (2003). Subjunctive Verbs: A Result of Natural Grammatical Processes in First Language Acquisition, Second Language Learning, Language Variation, and Language History. In The Development of Grammar in Spanish and The Romance Languages (pp. 70-90). Palgrave Macmillan, London. Ferrell, J. E. (1999). The development of subjunctive use with expressions of emotion from Latin to Spanish. University of Michigan. Floyd, M. (1990). DEVELOPMENT OF SUBJUNCTIVE MOOD IN CHILDREN'S SPANISH: A REVIEW. Confluencia, 5(2), 93-104. Retrieved July 28, 2021, from http://www.jstor.org/stable/27921930 Gallego, M., & Alonso-Marks, E. (2014). Subjunctive use variation among monolingual native speakers of Spanish: A cross-dialect analysis. Spanish in Context, 11(3), 357-380. George Bogdan Țâra. (2018). A perspective on the vocabulary common to Classical and Vulgar Latin. Diacronia, 7. https://doi-org.proxy.mtholyoke.edu:2443/10.17684/i7A98en Giancaspro, David. Heritage speakers' production and comprehension of lexically- and contextually-selected subjunctive mood morphology. Retrieved from https://doi.org/doi:10.7282/T3BG2S3N Godard, D. (2012). Indicative and subjunctive mood in complement clauses: from formal semantics to grammar writing. Empirical issues in syntax and semantics, 9, 129-148. Grandgent, C. H. (1907). An introduction to vulgar Latin. D.C. Heath & Co. https://ebookcentral.proquest.com/lib/mtholyoke/detail.action?docID=622555 http://www.jstor.org/stable/27921930 https://doi-org.proxy.mtholyoke.edu:2443/10.17684/i7A98en https://doi.org/doi:10.7282/T3BG2S3N 132 Gregory, A. E. (2001). A cognitive map of indicative and subjunctive mood use in Spanish. Pragmatics & cognition, 9(1), 99-133. Guajardo, G. (2021). Two subjunctives or three? A multimodel analysis of subjunctive tense variation in complement clauses in Spanish. International Journal of Corpus Linguistics. Gudmestad, A. (2006). L2 variation and the Spanish subjunctive: Linguistic features predicting mood selection. In Selected proceedings of the 7th Conference on the Acquisition of Spanish and Portuguese as First and Second Languages (pp. 170-184). Cascadilla Proceedings Project Somerville, MA. Gudmestad, A. (2013). Tense-Aspect Distinctions within the Subjunctive Mood in the Spanish of Native Speakers and Second-Language Learners. Studies in Hispanic and Lusophone Linguistics, 6(1), 3-36. https://doi.org/10.1515/shll-2013-1139 Herman, J., & Wright, R. (2000). Vulgar Latin. [electronic resource]. Pennsylvania State University Press. Jensen, F. (2017). The syntax of the old French subjunctive. Mouton. Jensen, F., & Lathrop, T. A. (2017). The Syntax of the Old Spanish Subjunctive. Mouton; Academia Publishing. Kempchinsky, P. (2009). What can the subjunctive disjoint reference effect tell us about the subjunctive? Lingua, 119(12), 1788-1810. Klausenburger, J. (2000). Grammaticalization: studies in Latin and Romance morphosyntax (Vol. 193). John Benjamins Publishing. Lakey, H. (2015). The Grammaticalization of Latin nē + Subjunctive Constructions. Journal of Latin Linguistics, 14(1), 65-100. https://doi.org/10.1515/joll-2015-0004 Leroy Maurice. (1939). Thomas (François). Recherches sur le subjonctif latin. Histoire et valeur des formes. Revue Belge de Philologie et d’histoire, 18(4), 993–996. Lindschouw, J. (2010). Grammaticalization and language comparison in the Romance mood system. In Modality and Mood in Romance (pp. 181-208). De Gruyter. https://doi.org/10.1515/shll-2013-1139 https://doi.org/10.1515/joll-2015-0004 133 Löfstedt, L. (2008). Le subjonctif imparfait de l'auxiliaire modal et l'infinitif du passé. Neuphilologische Mitteilungen, 109(2), 131-142. Retrieved August 3, 2021, from http://www.jstor.org/stable/43344349 Mackenzie, Ian. 1999–2022. The Linguistics of Spanish. http://www.staff.ncl.ac.uk/i.e.mackenzie/index.html (homepage). 20 April 2022. Mackenzie, I. E. (2019). Language Structure, Variation and Change. [electronic resource] : The Case of Old Spanish Syntax (1st ed. 2019.). Springer International Publishing. Márquez, R. R., & Placencia, M. E. (Eds.). (2004). Current trends in the pragmatics of spanish. ProQuest Ebook Central https://ebookcentral-proquest-com.proxy.mtholyoke.edu:2443 Meagan Ayer, Allen and Greenough’s New Latin Grammar for Schools and Colleges. Carlisle, Pennsylvania: Dickinson College Commentaries, 2014. ISBN: 978-1-947822-04-7. https://dcc.dickinson.edu/grammar/latin/credits-and-reuse Mejías-Bikandi, E. (1998). Pragmatic presupposition and old information in the use of the subjunctive mood in Spanish. Hispania, 941-948. Monachesi, P. (2005). The verbal complex in romance : A case study in grammatical interfaces. ProQuest Ebook Central https://ebookcentral.proquest.com Murphy, M. D. (2008). The role of typological drift in the development of the Romance subjunctive: a study in word-order change, grammaticalization and synthesis. The University of Texas at Austin. Panzeri, F. (2019). In the (Indicative or Subjunctive) Mood. Proceedings of Sinn Und Bedeutung, 7, 215-227. https://doi.org/10.18148/sub/2003.v7i0.804 Poplack, S., Cacoullos, R., Dion, N., Berlinck, R., Digesto, S., Lacasse, D. & Steuck, J. (2018). 8. Variation and grammaticalization in Romance: a cross-linguistic study of the subjunctive. In W. Ayres-Bennett & J. Carruthers (Ed.), Manual of Romance Sociolinguistics (pp. 217-252). Berlin, Boston: De Gruyter. https://doi.org/10.1515/9783110365955-009 Poplack, S., Lealess, A. & Dion, N. (2013). The evolving grammar of the French subjunctive. http://www.jstor.org/stable/43344349 https://ebookcentral-proquest-com.proxy.mtholyoke.edu:2443/ https://ebookcentral.proquest.com/ https://doi.org/10.18148/sub/2003.v7i0.804 https://doi.org/10.1515/9783110365955-009 134 International Journal of Latin and Romance Linguistics, 25(1), 139-195. https://doi.org/10.1515/probus-2013-0005 Portner, P., & Rubinstein, A. (2020). Desire, belief, and semantic composition: variation in mood selection with desire predicates. Natural Language Semantics, 28(4), 343-393. Reyes, Lucia, et al. “Language and Therapeutic Change: A Speech Acts Analysis.” Psychotherapy Research, vol. 18, no. 3, May 2008, pp. 355–362, 10.1080/10503300701576360. Accessed 10 Jan. 2022. Schwenter, S. A., & Hoff, M. (2020). Cross-dialectal productivity of the Spanish subjunctive in nominal clause complements. Variation and Evolution: Aspects of language contact and contrast across the Spanish-speaking world, 29, 11. Sessarego, C. (2016). A discourse-pragmatic approach to teaching indicative/subjunctive mood selection in the intermediate Spanish language class: New information versus reformulation. Hispania, 392-406. Shawl, J. (1975). Syntactic Aspects of the Spanish Subjunctive. Hispania, 58(2), 323-329. doi:10.2307/338959 Vesterinen, R. (2013). Instructions or dominion?: The meaning of the Spanish subjunctive mood. Pragmatics & cognition, 21(2), 359-379. Vesterinen, R., & Bylund, E. (2013). Towards a unified account of the Spanish subjunctive mood: Epistemic dominion and dominion of effective control. Lingua, 131, 179-198. Villalta, E. (2001). A comparative semantics for the subjunctive mood in Spanish. AMSTERDAM STUDIES IN THE THEORY AND HISTORY OF LINGUISTIC SCIENCE SERIES 4, 227-242. Villalta, E. (2008). Mood and gradability: an investigation of the subjunctive mood in Spanish. Linguistics and philosophy, 31(4), 467-522. Viner, K. M. (2018). The optional Spanish subjunctive mood grammar of New York City heritage bilinguals. Lingua, 210, 79-94. Whitney, W. D., and JSTOR. The Principle of Economy as a Phonetic Force. Internet Archive, https://doi.org/10.1515/probus-2013-0005 135 Transactions of the American Philological Association (1869-1896), 1 Jan. 1877, archive.org/details/jstor-2935726/page/n1/mode/1up. Accessed 8 May 2022. Winters, M. E. (2013). Grammatical meaning and the old French subjunctive. In Research on old French: The state of the art (pp. 351-376). Springer, Dordrecht. Zavadil, B. (2015). Historia de la lengua espaňola : Introducción a la etimología. ProQuest Ebook Central https://ebookcentral-proquest-com.proxy.mtholyoke.edu:2443 https://ebookcentral-proquest-com.proxy.mtholyoke.edu:2443/