THE FORMATION AND THE COGNITIVE KNOWLEDGE OF LITERARY LANGUAGES: THE CASE OF HEBREW AND ARAMAIC IN THE MIDDLE AGES

This paper characterizes Medieval Hebrew and Aramaic as literary languages and seeks to explain how a ‘literary language’ – namely a language used mainly in literary contexts – arises, while utilizing three types of research: comparative philological research, which compares different languages and texts in terms of their vocabulary and grammar; sociolinguistic research, which examines the social functions of language use; and psycholinguistic research, which (in this particular case) examines issues of language acquisition. The paper builds on philological studies of literary languages to explain how the grammar of these languages evolves. It assumes that the acquisition of such languages is similar to second-language acquisition, while taking into account that these languages are both acquired and used in a strictly literary context. The main argument of the paper is that literary languages should be studied the same way as other languages, because ultimately – after making some adjustments motivated by their particular functions – they are compatible with the standard models of second-language acquisition.


INTRODUCTION: THE CHARACTERIZATION OF MEDIEVAL HEBREW AND ARAMAIC AS LITERARY LANGUAGES
The Jewish legal and exegetical scholarship of the Middle Ages centered around the corpus of classical texts in Hebrew and Aramaic, from the Bible to the Tannaitic and Amoraitic literature. Most of the scholars of the period even wrote in Hebrewor, in the case of commentary on the Babylonian Talmud, in a language similar to that of the Babylonian Talmud itself, which combines Hebrew and Aramaic. 1 After the Aramaic seized from being the spoken language in most of the Jewish diaspora, very few works were written entirely in it, 2 with the exception of the Targum project, which was completed during this period. 3 The situation changed with the advent of the Jewish Kabbalah, whose main text, the Zohar, 4 was written mostly in Aramaic, giving rise to a literary dialect 5 with distinct linguistic features. In this paper I seek to focus on Hebrew and Aramaic in this period and to address the broad question of how a "literary language"a language used mainly in literary contextsis formed. 6 This study seeks to combine three types of research: a. Comparative philological researchwhich compares different languages and texts in terms of their vocabulary and grammar. b. Sociolinguistic researchwhich examines the social functions of language usage. c. Psycholinguistic researchwhich examines questions of language acquisition. 7 The study aims to utilize the findings of philological inquiries on languages whose socio-functional use is for literary purposes. Its goal is to determine how the grammar of these languages evolves, based on the premise that their acquisition is similar to the acquisition of a second language, but also while recognizing that these languages are both acquired and used in a strictly literary context. My aim is to shed light on how new vocabulary and new grammar evolve in the context of a literary language.
The main argument of this paper is that literary languages should be studied in the same way as other languages, because these languages are ultimately compatible with the standard models of second language acquisition, with some adjustments motivated by their particular function.
The structure of the paper is as follows: Section 2 examines three possible ways of regarding literary languages: (a) as dead or artificial languages; also Y. M. Mayer, "Crying at the Florence Baptistery Entrance -A Testimony of a Traveling Jew," Renaissance Studies 33 (2018), pp. 441-457. 6. While Biblical and Rabbinic Hebrew are literary languages in the sense that they are known to us only from texts, there are good reasons to believe that close variants of them were used as vernaculars at some point. However, it appears that during the Interim Period Hebrew was no longer spoken on a daily basis. , who characterized Rabbinic/Mishnaic Hebrew as a "living literary tongue." While the texts examined in this paper are medieval, in terms of the language we are interested in the Hebrew of the Interim Period, during which Hebrew was no longer in daily use as a spoken vernacular. Obviously, the Interim Period is not coextensive with the medieval period. 7. For a discussion on the need to include psycholinguistic and sociolinguistic aspects in discussion on historical linguistics, see L. (b) using the diglossic model and a contrastive analysis approach; (c) using the diglossic model but coupled with an approach that assumes the development of a "learner's language." I will argue for the adoption of the third option. Assessing options (b) and (c), I will present the psycholinguistic model and show how it can be relevant to the study of literary tongues; in footnotes I will outline how the classical philological study of medieval Hebrew and Aramaic relates to these issues. In Section 3, I will discuss some unique features of literary languages. Section 4 will provide an interim summary of all of the characterizations of literary languages discussed in this paper. Section 5 will elaborate briefly on some aspects of the grammar of literary languages associated with their functions. Section 6 concludes. Beyond this, since the claims of this paper are general, and presume to apply to literary languages other than the Jewish ones, I shall also draw parallels to the study of Latin and Sanskrit, which had a similar status for long periods of their history.
In the context of the current issue, this study is important to the discussion on the "revival" of Modern Hebrew in various ways. First, it provides a better understanding of the nature of the knowledge of Hebrew prior to the period when its use as spoken language restituted. Second, it is in fact still relevant for understanding the cognitive linguistic knowledge of the first authors who began to write in Hebrew as part of its ideological revival. 8 In addition, it may shed light on some of the process that Modern Hebrew had to go through such as the phenomenon of regularization. 9

THE CHARACTERISTICS OF LITERARY LANGUAGES 2.1 Literary Languages as Dead or Artificial Languages
Languages that are no longer spoken are standardly referred to as "dead" or "artificial." Thus, the Hebrew of the Interim Period, between Mishnaic and Modern Hebrew, has often been described as a "dead language." 10 This notion 8. See Stern, in this issue. 9. See Bar-Ziv Levy, in this issue. 10. On this see Rabin, The Development of the Syntax, pp. ix-x. It is noteworthy that although E. Goldenberg, "Hebrew Language. Medieval," in Encyclopedia Judaica (2 nd edition, Volume 8, Detroit, New York, San Francisco, New Haven, Waterville, London 2007), p. 650, states that medieval Hebrew "lived 'an active life' in written texts," she still distinguishes it from "living languages." The term "artificial," often applied to Medieval Hebrew and Aramaic, will be discussed below.
is still with us today, as evident from many descriptions of the history of Hebrew that refer to the "revival" of the language in the modern era. 11 This term implies that Hebrew was previously dead, and if it is applicable to Hebrew, it must necessarily be applicable to its lesser sister, Aramaic, as well. In discussions that take this approach, the criterion for describing a language as "dead" is the absence of living speakers, 12 and sometimes a stricter criterion is used: the absence of native speakers.
As already pointed out by scholars of Latin and Sanskrit, the terms "living" and "dead" as applied to language are vestiges of the organic perception of language that was dominant in the first half of the 19 th century. In this approach, articulated most explicitly by August Schleicher, language was viewed as a living organism, and it was often seen as independent of its speakers. 13 Since the time of the Neogrammarians of the late 19 th century, this view has been thoroughly rejected 14 in favor of a perception of language as a 11. E.M Lifschitz, ‫,כתבים‬ (Writings, Jerusalem 1941), p. 52 uses the terms "dead" and "living": "It has often been remarked that the condition of our language is a strange and unclear one, for it was simultaneously alive and dead throughout the Middle Ages. This, indeed, was its condition: it was a living literary tongue rather than a living spoken tongue. Even in the Middle Ages our language exhibited every characteristic of a living literary tongue: it served the needs of [successive] generations and was adapted to express all their thoughts; new expressions were often created in it, and new ways of speech to express their logic. It did not force its users to garb their thoughts in a borrowed mantle or in pre-existing florid idioms. That was the nature of our language as a literary tongue throughout the Middle Ages" (see also ibid., p. 252 for similar remarks). 12. Z. Ben-Hayyim, ‫לתקופות‬ ‫וחלוקתה‬ ‫העברית‬ ‫הלשון‬ ‫של‬ ‫ההיסטורית‬ ‫האחדות‬ -‫כיצד?‬ (The Historical Unity of Hebrew and its Periodical Division), in Studies in Language 1 (1986), pp. 3-25, for example, suggested that the only valid criterion for delineating periods in the history of Hebrew is whether the language was written or spoken. 13. In the introduction to this work which came to be the archetype for comparative linguistic research, Schleicher describes language as an organism, and maintains that the development of every language traces an arc: it gradually grows to attain its mature and perfect form, following which it begins to decline towards its ultimate death (A. Schleicher, Compendium der vergleichenden Grammatik der indogermanischen Sprachen (Kurzer Abriss der indogermanischen Ursprache, des Altindischen, Altiranischen, Altgriechischen, Altitalischen, Altkeltischen, Altslawischen, Litauischen und Altdeutschen, I-II, Weimar 1861-1862). This organic perception, which assumes that languages are ontological entities independent of their speakers, is explicitly articulated in his other treatise A. Schleicher, Die Darwinsche Theorie und die Sprachwissenschaft offenes Sendschreiben an Herrn Dr. Ernst Haeckel, (Weimar, 1863), which compares the basic assumptions of comparative linguistics to Darwin's theory of evolution (as Schleicher understood it). 14. Heller, for example, comes out against the tendency to speak of language as an organism. He argues that a language cannot be dead because it was never born, and that it is the speakers, rather than the tongue, who cognitive-psychological phenomenon. However, even the critics of the organic approach often continued using organic terminology in describing the history of language, 15 and this seems to have affected the discussion of languages that lived mainly as literary tongues.
Paradoxically, the cognitive approach had a similar effect as well, for it motivated an emphasis on "natural languages," and assumed that only native tongues are natural, whereas others are artificial. Generative research following Noam Chomsky, which sought to characterize innate linguistic knowledge, focused on speakers' first languages, 16 and did not consider second and literary languages to be valid objects of theoretical linguistic study. However, even if we accept Chomsky's claims about first language acquisition, literary languages still present a question for research, namely what linguistic knowledge is utilized by their users? If the knowledge underpinning the use of a second language is a legitimate object of research, literary languages can be explored from the same perspective. According to this approach, describing such languages as "artificial" contributes nothing to their theoretical understanding. In order to avoid such characterizations, even as metaphors, we must look for some other way to make an essential distinction between spoken languages and ones whose entire existence is in the written domain. Hence, in characterizing the various layers of Hebrew and Aramaic, we must answer the following questions: What essential change occurred during the Middle Ages as these languages became literary tongues? How does this change affect the linguistic features of these languages? How can the users' knowledge of literary languages be characterized?
In this paper, I wish to adopt the approach of David Téné (also adopted by scholars of Latin and Sanskrit in the relevant periods), who argued that the literary Hebrew of the Interim Period was in an inherent state of diglossia. 17 According to the model proposed by Charles Ferguson, 18 disglossia is a situation in which a speech community speaks or uses two languages simultaneously, 19 or else two systems of the same language. The term applies to situations where each system serves a different function, so that the two are in complementary distribution in terms of their use. In analyzing states of diglossia, Ferguson distinguishes between High (H) and low (L) language, the former being the marked case, reserved for special cultural contexts, and the latter being unmarked and used in everyday situations. According to this model, the Jewish authors of the medieval Hebrew and Aramaic texts functioned in a state of diglossia where their native vernaculars were often used only as low languages, whereas written Hebrew and Aramaic served as high languages, used mainly in contexts of study and scholarship but also as means of written communication. These languages were sometimes spoken as well, used as lingua francas, 20 but that is not our focus here, so I will not , for many testimonies about Hebrew spoken during the Middle Ages. His accounts are problematic, however, since he does not differentiate between solid evidence and myths. I thank Miri Bar-Ziv for pointing me to Haramati's book. discuss the issue of whether and how these languages were used as pidgins or as a Koiné. 21 The next section explores the implications of characterizing a language as a literary tongue used in a diglossic context. So far, I have mentioned only the sociolinguistic aspects of diglossia, namely its functional aspects. The section below examines its psycholingustic implications, i.e., what diglossia means for the acquisition of a literary tongue. As mentioned, this section will also examine how philological studies of the Hebrew of the Interim Periodi.e., from the time it ceased to be spoken in Late Antiquity until its revival as a vernacular in the Land of Israelcan shed light on the phenomenon of diglossia. In addition, I will point to parallels from Sanskrit and Latin in the relevant periods. In the case of Latin, this refers to the Middle Ages and the Renaissance, during which it served as the cultural language of Europe and thus existed as a literary language in a diglossic context; as for Sanskrit, all the texts known to us are apparently from the period when it was no longer a spoken tongue. 22

Literary Languages in Their Diglossic Context: The "Contrastive
Analysis" Approach Already in his first study that presented the phenomenon of diglossia, Ferguson noted that when languages are in a state of diglossia in terms of their distribution, the H variety is acquired as a second language. 23 Children are obviously exposed to the L variety first, and only later master the H varieties, especially at an advanced level of proficiency. It therefore seems pertinent to examine the implications of this insight for the analysis of literary languages, in light of studies that explored the unique aspects of second language acquisition. 24 It should be kept in mind, however, that in the case of literary languages in a state of diglossia, the second, "acquired" language is not the native tongue of any speech community, and its learners are exposed to it only in its literary uses.
To facilitate the discussion, let me first of all present some terminology: The study of second language acquisition (SLA), from either a psychological or a pedagogical perspective, seeks to identify the factors that determine how successful a learner will be in acquiring the target language, based on the understanding that this kind of acquisition differs from a child's acquisition of his first language. As a matter of fact, after puberty flawless acquisition of a second language is impossible, so that the grammar of the learner will necessarily differ from that of a native speaker.
Until the 1970s it was accepted that limitations on SLA are the result of the learner's established knowledge of her native tongue, which interferes with the learning of a second grammar. Accordingly, it was assumed that every discrepancy between the grammar of a learner and that of a native speaker can be explained as a reflection of the learner's native tongue. This approach, presented by Robert Lado 25 and termed "contrastive analysis," applies knowledge gathered in studies of "languages in contact," which examine how the grammar of one language influences the grammar of another when both languages are in simultaneous use by the same speakers. Studies of languages in contact, such as Uriel Weinreich, 26 were indeed central to this approach to SLA.
Applying this model to SLA in contexts of diglossia involves identifying instances in which the grammar of the learner's first language influences the grammar she constructs for the target language. 27 Such influence is expressed first of all in pronunciation, but also in syntactic structures and of course in the lexicon (the borrowing of words and of meanings). All of these are transfers, overt or covert, from the first language to the second, and the claim is that diglossic situations, in which the high language is acquired as a second language, involves similar processes. According to this model, the acquisition of a literary language, like the learning of any second language, involves an attempt to internalize all the aspects of the target language (to be discussed below), but results in a language that differs from the target language in some respects, because it incorporates specific phenomena originating in the learner's native tongue. In other words, studying the syntax and vocabulary of the target language as reflected in the various texts results in "transfer", namely in the duplication and internalizing of its structures and forms, to the point of achieving a degree of mastery. However, this process is never perfect, due to the tendency to transfer not only from the target language but also from one's first language. Hence, a form internalized by the learner may accurately mirror the target language as reflected in the texts, or may also be taken from the learner's native spoken tongue (lexical borrowing). The influence of the first tongue is sometimes reflected only when comparing the meanings of a certain expression (cases of calques), either at the level of words or at the level of syntactic constructions, manifested in formal similarities (of word order, agreement, definiteness, etc.).
As a matter of fact, much of the study of Medieval Hebrew has focused on this issue, namely on examining Hebrew texts for the influence of the local vernaculars. Observations of this sort feature in all types of linguistic analyses of Medieval, enlightenment-period and revival-era Hebrew. 28 The influence of local vernaculars was similarly addressed by researchers of Neo-Latin. 29 However, as we shall see, the contrastive analysis approach is no longer accepted in psycholinguistic research. In the next section I will present the approach that has superseded it, following which I shall examine the implications of the latter approach for the study of literary languages.

Characterizing Literary Languages in Their Diglossic Context: The
Approach Focusing on the Development of a "Learner's Language" Researchers began questioning the contrastive analysis approach in the early 1970s. Pit Corder's studies focused instead on "errors" in the grammar of second-language learners, i.e., places where their grammar differs from that of native speakers (of the target language). 30 He argued that not all these differences can be explained in terms of transfer from the learner's first language; some stem from other reasons, having to do with innate linguistic knowledge or with the process of learning and internalizing the target grammar. A seminal study of these issues was published by Larry Selinker, 31 who developed the idea of an interlanguage that is intermediate between the learner's native tongue and the target language. Although the learner aspires to master the target language, the grammar she constructs in the process of learning is also influenced by her general learning strategies and strategies of acquiring and communicating in the new language. The important insight, from our perspective, is that the process of acquisition involves the construction of grammatical rules that do not exist either in the target language nor in the learner's native speech. Moreover, Selinker argues that learners often reach a state of "fossilization," at which some of these rules become permanent, thus preventing the learner from attaining a perfect mastery of the target language.
Important for our purposes is the observation that the learning process often involves overgeneralization, which is also detected in first language acquisition (FLA). Children acquiring English, for example, often overgeneralize the plural suffix, applying the suffix -s to nouns such as mouse, thus producing the form mouses instead of the correct irregular form mice. In morphological contexts, overgeneralization stems from the application of morphological rules based on analogy. Following de Saussure, structuralist linguists hold that the acquisition of grammar, at least at the morpho-lexical level, depends on the learner's ability to draw analogies and incorporate them into her linguistic knowledge, and that this ability is a basic component of language acquisition in children. 32 Most overgeneralizations involve the application of such analogies to "irregulars," as in the example presented above. But the mechanism of misidentifying and misapplying a rule is in fact broader, and can lead to the construction of new morphological paradigms. Furthermore, overgeneralization is not confined to morphology, but is relevant to syntax as well.
The important insight of Selinker and others is that overgeneralization occurs not only in FLA but also in SLA. 33 That is, an adult learning English as a second language may also construct a grammar that yields the form mouses (and, if it becomes fossilized, this can become a permanent feature of his English). It goes without saying that the same phenomena are also responsible for changes that languages undergo over time and are traditionally explained in terms of analogy.

Applying the "Learner's Language" Approach to Literary
Languages Are the phenomena discussed in the previous section also relevant to literary languages? As stated above, studies of SLA rejected the approach of contrastive analysis, which assumes that the learner's language is identical to the target language except where transfer from her native tongue occurs, and adopted a different view: that the process of learning involves the creation of a new grammar whose features cannot always be explained in terms of transfer from the first language. I argue, in a similar vein, that literary languages in diglossic contexts are not just classical languages with errors introduced under the influence of the vernacular. Rather, they too undergo changes that include the creation of new grammars.
Following this line of thought, I reject the belief dominant among researchers, including researchers of Hebrew, that analogies generating new grammar occur only in spoken languages, 34 because this belief is not supported empirically, as several researchers have already demonstrated with respect to Hebrew. 35 The same has been shown with respect to other literary languages, 36 as I discussed extensively in a different context. 37 As stated, these analogies involve the same mechanism as overgeneralization. These phenomena are also familiar in the context of SLA, and many scholars have noted that they are especially prevalent in contexts of diglossia. 38 Accordingly, I claim that one of the most fascinating aspects in the study of literary languages is the identification of new grammatical rules, morphological and syntactic.
We can thus identify another motivation for examining the diglossic status of literary languages in light of insights from SLA studies. Just as the "errors" made by learners of a spoken second language shed light on the linguistic knowledge unique to their interlanguage, examining the "errors" in literary languagesthe places where they differ from the original languagecan shed light on the linguistic knowledge of their users, who employed it in the written medium. It is important to stress that, in the case of literary languages, the entire language is in a "fossilized" state, and there are no native speakers to whom the learners can be compared.
While grammatical innovations have not been the focus of the inquiry into literary languages, neologisms (lexical innovations) have been more widely who made similar arguments regarding the printing of the Mishna in Italy. Observations of this sort were in fact made even earlier. Rabin, )The Development of the Syntax, pp. 89-92) observed that in the Medieval Hebrew of Spain there was a tendency to determine the grammatical gender of nouns according to their suffix, so that only nouns ending in -a were regarded as feminine. He also noted the widening of phenomena that existed in ancient Hebrew, and the creation of new patterns that had not existed before (p. 168). Languages, (Jerusalem, 1970) discusses numerous cases of languages in diglossic distribution where the high variety exhibits numerous hypercorrections; these stem from differences between the spoken and high varieties and are motivated by a desire to "purge" the high language of these differences even in cases where its grammar does not actually require it. In the first chapter of his book, Blau presents principled differences between this phenomenon and regular analogies, but acknowledges that at their basis they involve similar mechanisms. For a review of the phenomenon of hypercorrections in Sanskrit see G. M. Lee, "Diglossia in Ancient India," Working Papers in Linguistics 34 (1986), pp. 155-160. noted. 39 In this respect, I discern a certain difference between the learners of a second language that is spoken and the learners of a literary language. The former do not usually create new lexical items in their second language. This presumably stems from a sense of "foreignness": speakers of a second language do not regard themselves as equal members of the speech community who have the right to innovate. This feeling is absent in the diglossic situation where the second language is a literary one. In this situation, users have a sense of "cultural ownership" over the language, and lexical innovations are therefore common; moreover, they are necessary in order to preserve the relevance of the language 40 and are unavoidable if the language is used to describe new ideas and discoveries. We may thus be witnessing the influence of the sociolinguistic situation (status) on linguistic performance (the users' inclination to take part in expanding the lexicon).
In sum, innovations, in the form of neologisms as well as new grammars, reveal the linguistic knowledge of the users of literary languages, 41 knowledge which in many ways resembles that of learners of spoken languages.

DIFFERENCES BETWEEN THE ACQUISITION OF A LITERARY
LANGUAGE AND AN ORDINARY SLA: THE MULTIPLICITY OF GRAMMARS Alongside the similarity, it is important to note some differences between the acquisition of a high literary language and the acquisition of an ordinary second language. These differences once again highlight the importance of the sociolinguistic context, namely the fact that these languages are acquired as literary tongues.
Those who learn literary languages are often exposed to a range of different sources in these tongues, from different periods and different dialects, thus exhibiting different grammars. In this situation, learners cannot be expected to distinguish between various diachronic layers in the texts, 42 and distinct grammars may also develop due to different levels of exposure to different texts or due to different ways of resolving grammatical discrepancies among the classical texts. 43 In the case of Hebrew and Aramaic, those who learned Hebrew and Aramaic from the classical texts during the Middle Ages read texts from different sources in these tongues, 44 from different periods, which clearly exhibit different grammars. 45 These sources included the Bible and its Aramaic translation; the Tannaitic Hebrew literature; the Amoraitic literature, which combines the two languages, and the medieval literature written up until their timeas well the Hebrew and Aramaic that were being used in their environment, mostly in writing. 46 Sometimes, however, we encounter conscious decisions regarding what should be incorporated in the grammar and what should not be. Thus, it is possible to identify choices between rival grammars in medieval Hebrew, especially in the choice between Biblical and Mishnaic Hebrew. 47 A similar phenomenon is observed in other literary languages, for example in the debate among the Humanists on whether or not Cicero's language is the only language worthy of emulation. The division into ages and the value judgments about them (as reflected by the terms "golden," "silver" and "copper" ages) played a central role in this debate. 48 In addition, we must mention not only cases of competing grammars of the same language, but also the phenomenon of existence of parallel literary languages (triglossia). Latin and Greek in the Indo-European world 49 and Hebrew and Aramaic in the Jewish context represent a situation where two literary languages existed side by side, both of them belonging to the same family and possessing relatively similar grammar. This meant that the grammar of one could influence the grammar of the other, 50 for example

AN INTERIM SUMMARY
Diagram 1 summarizes what has been said so far in this paper about the sources of the linguistic, grammatical and lexical competence underpinning the use of literary languages.
Diagram 1: the sources for the grammar of literary languages According to this description, the essential differences between the acquisition of a (second) literary language and a (second) spoken language stem from differences in the process of acquisition. The facts reviewed above show that there are differences both in usage and in the nature of the knowledge acquired. The complex picture I have painted depicts a state of Hebrew demonstratives; and Goldenberg "Hebrew Language. Medieval," p. 669, on Aramaic words in Sefer Hasidim. 51. The present paper focuses on the acquisition of the target language (Medieval Hebrew and Aramaic), and therefore does not discuss the influence of the target language (specifically Hebrew) on the user's native tonguean issue that is in fact the focus of most studies of "Jewish languages," which extensively discuss the "Hebrew component" in Jewish languages. For a recent discussion, see M. Bar-Asher, ‫חיה"‬ ‫כלשון‬ ‫העברית‬ ‫היהודים‬ ‫לשונות‬ ‫"בתוך‬ (Hebrew as a Living Language in Jewish Languages), in Carmillim 13 (2017-2018), pp. 9-26, on the vitality of Hebrew within the Jewish tongues and the semantic and morphological development of forms. instability and competition between parallel grammars, and suggests that, contra to what might be expected, the linguistic diversity that exists in every language 52 is actually intensified in literary languages, and this is the background for the need for "regularization" during the process of the emergence of Modern Hebrew. 53 It would be highly interesting, I believe, to characterize this kind of knowledge from a psycho-linguistic perspective, but this enterprise is beyond the scope of the present study.

CHARACTERIZATIONS RELATED TO THE FUNCTION OF LITERARY
LANGUAGES Alongside the similarity between first-language and second-language acquisition, it is important to note what distinguishes the acquisition of high or literary languages; that is, we now return to discuss the unique function of literary language. While studies of SLA focus on the acquisition of language for purposes of everyday communication, our focus is on languages whose sociolinguistic function is in the literary-cultural domain.
SLA studies generally distinguish between "learning" and "acquisition," or between "explicit" and "implicit" acquisition. 54 This distinction is relevant to our discussion as well. Classic studies of diglossia generally assume that the learning of the second language involves a component of literacy, i.e., the systematic, formal learning of reading and writing, and also of grammar and stylistics. This is the approach taken in studies of Classical Arabic, for example. However, literary languages are not necessarily acquired only in this manner. A high language, just like any other second language, can also be learned through usage, in the form of exposure to educated speech and apparently even exposure to texts. 55 Despite this, we are still speaking of languages that are known from texts and used almost exclusively for literary purposes.
Apparently, such was the situation of Hebrew and Aramaic, 56 as well as of other classical languages that were used in intellectual and religious contexts. This can have significant implications, because certain kinds of learning may give rise to particular forms of knowledge and not to others. For example, only a particular kind of learning can provide users with the degree of expertise needed to distinguish between different ancient layers of the tongue and employ a puristic strategy that favors one layer over the other.
Another important point relevant to the acquisition of language is the following: Psychological studies have shown that, in both SLA and FLA, learners tend, at least in the initial stages of learning, to acquire expressions as a whole, without analyzing them into their component parts (i.e., in a noncompositional manner). For example, a student of Hebrew may learn to say ma shlomxa ("how are you", lit. what [is] your wellbeing) as a polite greeting, without recognizing-xa as the second-person possessive suffix and being able to use it in other contexts, for example to create the combination yalde-xa ("your boy"), especially since such pronominal suffixes are not productive in 55. On this issue with regard to Sanskrit, see Houben, "Socio-linguistic Attitudes," pp. 178-179, and especially M.M Deshpande, in Sanskrit & Parakirit: Sociolinguistics Issues, Delhi 1993, pp. 30-31, who stresses that Sankrit was learned through usage, although not necessarily in everyday contexts. Helander, "Neo-Latin," pp. 37-38 makes a similar claim regarding Latin, stating that in the Neo-Latin period, learned individuals were constantly exposed to this language. They heard it in lectures and spent much time reading texts in it. Hence, most of their learning was through imitation rather than the formal study of grammar. In the 17 th century Humanists were exposed to Latin in numerous ways; some learned mainly through imitation while others learned grammar more systematically (see Benner and Tengström, "On the interpretation," pp. 7-8, for a discussion of this, as well as a comparison to the situation in the Middle Ages). 56. Although grammarians regarded the grammatical knowledge of language mostly as a useful pedagogical tool, to my knowledge no study has examined how extensively it was used in teaching, nor has any study focused on the teaching of Hebrew and Aramaic in the Middle Ages. F.E Talmage, "Keep Your Sons from Scripture: The Bible in Medieval Jewish Scholarship and Spirituality," in Understanding Scripture, ed. C. Thoma and M. Wyschogrod, (New York 1987), pp. 81-101, makes some reference to this issue when he points to a difference between Ashkenaz and Spain in terms of the importance accorded to the teaching of grammar in the study of the Bible. Largely under his influence, Kanarfogel makes some comments on the issue in his discussions of Jewish education in the Middle Ages, and also provides some references to literature on specific aspects (E. Kanarfogel, Jewish Education and Society in the High Middle Ages, (Detroit, Michigan, 1992), pp. 31, 79-83). spoken Hebrew. In such cases the prevalence of the phrase is highly relevant: in both SLA and FLA, it is common phrases that tend to be learned in this manner. 57 This phenomenon occurs most naturally when language is acquired by hearing (or reading) it in use. As I demonstrate elsewhere, 58 it may have important implications for the analysis of literary languages as well.

CONCLUSIONS
The main argument of this paper is that, from a psycholinguistic perspective, the knowledge of literary languages can be investigated in the same way as the knowledge of any second language; however, attention must be paid to the sociolinguistic context, namely to the fact that literary languages are acquired through exposure to cultural textual sources and are used for literary purposes, usuallybut not necessarilyfor reading and writing texts in genres similar to those of the sources.