Loss and Consequence: An Examination of the Old English Case Marking System As Opposed to that of Other Old Germanic Languages Loss and Consequence: An Examination of the Old English Case Marking System As Opposed to that of Other Old Germanic Languages* * Denise E. Walters University of Colorado at Boulder Old English—early in its existence—did not differ much from its Germanic cousins. In fact, these languages could be considered distant dialects from one another. However, through the course of its development, Old English lost part of its Germanic morphology: the case-marking system. The loss of this system had such an impact on the development of the language that the results are seen in Modern English. This paper examines these results, and to an extent the reasons, behind this reduction. Introduction “Linguistic change is initiated by speakers, not by languages” (Milroy 1997: 311). Most linguists understand the truth behind this statement. One must take into consideration the speakers’ usage when considering language changes, regardless of the examination being undertaken. In the case of a diachronic study—even, or maybe especially, a typological one—the scholar must not only deal with the data, but also the relationship between that language and its cousins. Otherwise, a complete understanding might not be reached. For example, for the purposes of this study, Old English will be analyzed with relation to its Germanic cousins or sisters. Extensive changes in morphology occurred in late Old English (OE) and in Early Middle English (EME). Yet for various reasons, these changes did not occur within English’s Germanic cousins. The most obvious change was that the old system of case marking was nearly completely swept away. This drastic reduction of case marking has usually been seen as responsible for many syntactic changes, as well. These reductions are quite visible, as are the consequences of these reductions. Before the reductions occurred, English—in its infancy—looked so similar to its Germanic cousins that while the languages could not be considered dialects or variations1 of the same language, they were definitely not mutually unintelligible. 1. The Background Before examining the grammar of the nouns in OE that later underwent a change in case assignment and/or grammatical relations, it is necessary to make some general observations about OE syntax. In OE, case marking played an important role in signaling * I would like to thank Regina Pustet for her encouragement during her typology course. 1. For length reasons, these issues will not be addressed in this paper. Colorado Research in Linguistics. June 2004. Volume 17, Issue 1. Boulder: University of Colorado. © 2004 by Denise Walters. 1 Walters: Loss and Consequence Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 2 grammatical relations.2 Four cases were productive in OE: nominative, accusative, dative, and genitive.3 The latter three cases could all be used without a preposition to mark objects of verbs. The accusative was always used with verbs inherently high in transitivity (Andrews 1985); however, case marking is harder to predict with verbs that Hopper and Thompson (1980) consider less transitive. Moreover, many verbs show variability in their case marking. Most verbs which could take genitive objects also sometimes appear with objects in another case; the alternation between genitive and dative, as with gehelpan ‘to help’, is less common than that of accusative and genitive, as with afandian ‘to test, prove’ and abidan ‘to wait for, await’ (Bean 1983: 37). Afandian usually took a genitive object when it meant ‘test’ and an accusative object when it meant ‘prove.’ However, the accusative case was sometimes extended to the ‘test’ meaning in examples where it seems very difficult to argue for a difference in meaning which would explain the accusative (Bean 1983: 39). It is likewise with abidan: ‘to wait for’ requires the dative and ‘to await’, although semantically similar, requires the accusative (Bean 1983: 39). It is generally assumed amongst scholars that at least some case markings must be lexically specified, although these lexical specifications may well follow certain patterns based on semantics. Most current syntactic theories assume the existence of two distinct types of case marking: lexical (or inherent) and structural (or syntactic) case marking. Structural case marking is the default while lexical case is assigned idiosyncratically in lexical entries.4 Lexical case differs from syntactic case in that syntactic processes do not affect it. For example consider the difference between (1) a and (1) b in Icelandic: (1) a. Strákarnir voru kitlaðir boys-the-NOM-PL were tickled-NOM-MASC-PL ‘The boys were tickled.’ b. Stráunum var bjargað boy-the-DAT-PL was-SG rescued-NOM-MASC-SG ‘The boys were rescued.’ (Helfenstein 1870: 281) These sentences differ both in the surface case of the boys and in the agreement or lack of it of the passive participle and the auxiliary verb. The most widely accepted explanation for this difference is that the boys receives case structurally with kitla, but gets its case marking lexically from bjarga, which requires a dative object (Helfenstein 2. It must be noted, however, that even at the OE stage, a good deal of syncretism had crept into the case- marking system, due mainly to phonological change. For example, by the OE stage, the distinction between the nominative and accusative forms had been lost in a large class of nouns, the masculine a-stem, in both the singular and the plural (Allen 1995: 25). 3. Additionally, there was an instrumental case, but within prose, it had been almost completely replaced by prepositions, except within certain expressions. The instrumental was mainly retained in poetry and religious texts. 4. Precisely how structural case marking works will depend on the theory adopted. For example, Koopman and Sportiche (1990) suggest that there is more than one type of structural case marking; however, that will not be discussed here. It was merely mentioned to explain that different theories support different claims. 2 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/7 DOI: https://doi.org/10.25810/k43m-zz07 Loss and Consequence 3 1870: 282). The lexical entry of kitla does not impose any case marking on the arguments associated with the verb, and so these arguments get their case marking structurally based on their syntactic role at the surface level. OE was similar to Modern Icelandic in having a difference between lexical and structural case, although the case-marking patterns of the languages differ in some interesting respects, mainly due to the different principles followed in the application of case markings. However, at this point, it should be noted that dative and genitive case were regularly preserved under passivization in OE, as they are in Icelandic. For example, deman ‘to judge’ takes a dative object and the dative case remains when the verb is passivized: (2) hi ne demað nanum men, ac him they not judge no-PL-DAT men-PL-DAT but them-DAT bið gedemed is-SG judged. ‘They will not judge any men, but they will be judged.’ (Ælc.P.XI.369) However, the derived subjects of verbs which take accusative objects show up in the nominative case. The simplest explanation is that deman undergoes passivization just like verbs taking accusative objects, but because the underlying object of the verb receives dative case lexically, it does not receive nominative case structurally when becomes the subject. Being a case-marking language, OE signaled its grammatical relations by inflection, rather than constituent order, making possible much greater variation in the order of constituents than is found in the modern language. However, OE constituent order was not by any means completely free, and case marking was far from unambiguous. It is widely agreed that discourse factors played a very important role in OE constituent order; for example, most scholars concur that a noun phrase which had been mentioned before, and thus was ‘old information’ was likely to be placed near the front of a sentence.5 There is less agreement on the question of what role, if any, syntactic categories such as subject, verb, and object played. The most common argument is that both grammatical categories and discourse factors played an important role in OE constituent order, at least by the late OE period, as represented by Ælfric. There is one final aspect of OE syntax that must be talked about for background purposes: the use of the formal subject hit ‘it’. In contrast to Modern English (ModE), it is possible in OE for tensed clauses to appear with no NP in the nominative case. It is not difficult to find examples in OE in which a verb appears with a sentential complement and there is no anticipatory formal subject, which would be obligatory in ModE: (3) Đa gelamp þæt he… then happened that he ‘Then it happened that he…’ (Bede qtd. in Healey and Venezky 1980: 232). 5. Helfesntein 1870; Healey and Venezky 1980; Gneuss 1996 3 Walters: Loss and Consequence Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 4 It has been argued6 that formal hit was not used in the above sentence because its purpose in OE was simply to preserve verb-second order, and it was, therefore, not necessary when an adverb appeared before the verb. However, Allen (1986a) determined that this formal subject appears frequently in sentence in which it is not necessary for maintaining verb-second order. It is quite easy to find examples like (4), in which the verb would be in second position even with the addition of hit, and like (5), where its presence ensures that the verb is in third,7 rather than second, position: (4) þa gelamp hit þæt æt ðam gyftum… then happened it that at the wedding ‘Then it happened that at the wedding…’ (Ælc.TH.p.569) (5) On ðære tide iu hit getimode swa,… þæt he stod… in the time before it happened so,… that he stood… ‘At the earlier time it happened so…that he stood…’ (Ælc.P.XIV.I) A concordance corpus count of the verb gelimpan ‘to happen’ used with a sentential complement shows that the formal subject hit in fact appears in 83 percent of the examples in which it is ‘not necessary’ to maintain verb-second order, as in (4) and (5). Of the examples in which the formal subject is used, 71 percent do not have verb-second order (Allen 1986a: 468). It does appear, however, that hit was used to prevent verb-first order, since the placeholder is nearly always used when the verb would otherwise have been initial. Thus, formal subjects were greatly preferred in this sort of sentence in OE, regardless of whether the verb was in second or further position. The only positional constraint which played a role in the use of hit was the general—but not total—prohibition against verb-first declarative sentences. 2. The Case Markings of Old English: Explanation and Reduction8 Having established the background that informed the morphological changes, and therefore the syntactic changes, of OE, it is now time to talk about the changes themselves. The loss of case-marking distinctions in English has generally been seen as responsible for profound changes in the language. Compared with the present-day language, OE is highly inflectional. Nouns have four cases and three genders [cf. Appendix A for a brief summary of OE declensions]; verbs inflect for person and number and for the indicative and subjunctive moods. Further, in 6. Haiman 1974 7. That is, assuming that hit is to be treated as a separate constituent here, rather than as a clitic. If it treated as a clitic, then the verb is still in second position. 8. For purposes of this paper, only the case marking system of nouns and adjectives are to be considered. 4 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/7 DOI: https://doi.org/10.25810/k43m-zz07 Loss and Consequence 5 the OE noun phrase, there is agreement between noun and modifying adjective, much like the present-day German. The OE inflectional system derives directly from that in Germanic. However, OE begins to show the loss and simplification of inflections which characterizes the later stages of English and which eventually creates a language with remarkably few inflections compared to its cousins. One change which is frequently regarded as important in the loss of the ‘impersonal’ constructions is the syncretism between nominative and dative nominal cases. In OE, the distinctive dative case marking on the nominal clearly denotes which word is the ‘impersonal’ verb (6)9, but once the dative case marking disappears from the nominal paradigm, it becomes impossible for language-learners to be certain if the nominal present should be analyzed as an object (albeit, an indirect one) or as a subject. (6) þam cyninge ofhreoweþ the king-DAT to feel pity for something ‘The king pities…’ (Walters) Because the experiencer was in the preverbal position, which is typically occupied by the subject, the former object has been reanalyzed as a subject, despite the existence of examples with pronouns (7), in which the case marking is unambiguously non- nominative: (7) him ofhreoweþ he-DAT to feel pity for something ‘he pities…’ (Walters)10 The loss of the dative inflection for nouns was an instance of syncretism of forms, rather than loss of a category distinction, since the distinction between nominative and object case is still found in the pronouns. However, another change which has taken place, the collapse of the distinction between accusative and dative cases, involves the loss of an important category distinction, (8) and (9). The loss of this category distinction profoundly affects the case marking system of OE: the distinction between a lexically assigned case and a structurally assigned case is wiped out. (8) and him gelicade hire þeawas and þancode Gode and him like her virtues-NOM/ACC and thanked God ‘Her virtues pleased him, and he thanked God…’ 9. The nominal under discussion is underlined, and the verb is bolded. 10. The dative/nominative syncretism which occurred in the nominal system has also been regarded as the trigger for another sweeping syntactic change: the introduction of a new passive. For length reasons, this will not be discussed in this paper [cf. Allen, C. 1995. Case Marking and Reanalysis: Grammatical Relations from Old to Early Modern English. Oxford: Clarendon Press. ] The loss of this distinction is supposed to have led to unclarity about the grammatical relations involved in passives such as the king was given a gift and the king was harmed. In both of these passives, the king would have been clearly dative in OE, but was liable to reanalysis as the nominative subject at a later stage. 5 Walters: Loss and Consequence Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 6 or ‘He liked her virtues, and thanked God…’ (Healey and Venezky 1980: 352) (9) þonne soðilce Gode licað ure drohtnunge, þonne we þa then truly God-DAT likes our living, when we the god, þe we onginnað, on urhwuniendum end gefyllað good, which we begin, in preserving end fulfill ‘Then truly does our way of life please God, when we carry through to the end the good which we have begun’ or ‘Then truly does God like our way of life when we…’ (Healey and Venezky 1980: 120). As in most Germanic languages—or even languages from other branches of Indo- European—OE has four major types of vocalic nouns (nouns with vowels at the end of the stems): the a-stems, the ō-stems, the i-stems, and the u-stems, of which the first two are by far the most common. The a-stems are frequently referred to as the ‘masculine a- stems’, whose typical paradigm is represented by Table 1. Table 1. Declension of the masculine a-stems in OE. Example: stān ‘stone’ Singular Plural Nominative (NOM) stān stānas Accusative (ACC) stān stānas Genitive (GEN) stānes stāna Dative (DAT) stāne stānum The a-stems provide some good examples of how even at the earliest recorded stage of English, considerable syncretism of form—compared to the forms which can be reconstructed for Proto-Germanic—have taken place in English. For example, the nominative and accusative singular forms are distinct in the proto-Germanic language— and even in some of the other Germanic languages—but have fallen together by the earliest OE stage. The Germanic nominative singular form ended in –az, and the accusative form ended in –am. Both these forms disappeared by purely phonological processes which affected unaccented syllables in pre-OE [cf. Campbell 1959: 570], so that the nominative and accusative form for ‘stone’ converged on stān, with no suffix [cf. Icelandic harmr ‘harm’, in which z>r, but OE harm]. Syncretism had taken place in the plural even before the OE stage, with the original –ôs ending of the nominative extended to the accusative by the West Germanic stage (c.1100). Syncretism is also found in a group of nouns known as the ‘feminine ō-stems’, but the categories affected are different. In OE, the accusative, genitive, and dative forms of these nouns all ended in –e. This suffix is the reflex of three separate suffixes at the Germanic stage (Campbell 1959: 586), as shown in Table 2, which for length reasons only shows the singular paradigm. 6 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/7 DOI: https://doi.org/10.25810/k43m-zz07 Loss and Consequence 7 Table 2. Declension of the feminine ō-stems in OE. Example: giefu ‘gift’ (singular only) Proto-Germanic form of suffix NOM gief-u -ō ACC gief-e -ōm GEN gief-e -ôz DAT gief-e -ai Note: If stem is long, the –u of the nominative singular is deleted, as in lār ‘lore, doctrine’. Campbell indicates that the development to –e is regular by phonological processes, except in the genitive, where paradigmatic pressure and analogy seem to have played a role (1959: 586). Non-phonological pressures also play a role in the syncretism which took place in the nominative and accusative plurals of the feminine ō-stems. In Germanic, these suffixes were –ôz (NOM) and –ōns (ACC) (Lightfoot 2002: 99). The reflexes of these are –a and –e, respectively. One of the phonological processes contributing to much of the syncretism that took place by the end of the OE period was the reduction in the variety of vowels found in final unstressed syllables. The distinction between the back vowels in this environment was already showing clear signs of weakening in the Kentish charters of the 9th Century, and was completed in Northumbrian in the 10th Century, according to Campbell (1959: 377). The vowels seem to have remained distinct in West-Saxon for a longer period, although confusion of /a/ and /o/ is also apparent from scribal errors in early West-Saxon (Lightfoot 2002: 95). The front vowels were also affected, with /æ/, /e/, and /i/ falling together in a symbol written as at an early date (Campbell 1959: 369). This means that by the late OE stage, the only distinction in the vowels of suffixes was between higher and lower vowels. This distinction disappeared in the 11th Century, when the front and back vowels had ‘largely coalesced’ (Campbell 1959: 379). This reduction of unstressed vowels meant that the case inflections were now less effective than they had been at reflecting category distinctions. None of the above mentioned changes resulted in the loss of a category distinction. For example, although with many nouns the nominative and accusative are identical in form, the category distinction between nominative and accusative is still prominent, being reflected in the forms of adjectives and demonstratives, as well as in the forms of the feminine nouns [Appendix B. Tables 7 and 8, respectively]. Towards the end of the OE period, two phonological changes combine to have a devastating effect on the inflection of adjectives and determiners, as well as affecting the nominal declensions. The first change is a replacement of /m/ by /n/. The timing of this change with respect to other changes characteristic of EME has been documented by Moore (1928: 242). On the basis of an examination of a large number of texts, Moore concludes that this change was certainly completed by the end of the 11th Century (1928: 261). According to Campbell (1959: 378), the change is already evident in early West- Saxon in the dative plural ending of nouns, which is occasionally found as –un rather than the expected –um by the beginning of the 11th Century. 7 Walters: Loss and Consequence Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 8 The above change is by itself enough to cause a considerable amount of syncretism, especially since it was combined with the reduction of the vowels. For example, the dative plural –um was no longer distinct from the –an ending which was so frequent in the weak forms of the adjectives and also in a class of nouns, called ‘weak nouns’, which declined similarly to these adjectives. The syncretism becomes massive when another change quickly follows: the loss of final /n/ in unstressed syllables.11 This change affects the newly created /n/ as well as the old ones. For example, there is no longer any distinction between the nominative and accusative singular of any masculine nouns. This distinction has already disappeared in the strong a-stems, but in OE the distinction is maintained in the weak nouns (10). (10) hunta ‘hunter’ (NOM) versus huntant (ACC) (Walters) Earlier, the feminine weak nouns have had a distinction between –e in the nominative and –an in the accusative, but the endings combine when the nasal was lost and the vowels became identical. This means that some strong feminine nouns are the only ones which maintained the distinction between nominative and accusative singular (11). (11) dæ”d ‘deed’ (NOM) versus dæ”de (ACC) (Walters) It seems likely that it was impossible to maintain the distinction between nominative and accusative feminine nouns once so few nouns show this distinction (Moore 1928: 262). By the late 11th Century, the –e of all the non-nominative singular forms of these feminine nouns is extended to the nominative. With the reduction of final vowels and the loss of the final nasal, these nouns become in effect indeclinable, and the formal distinction between nominative and accusative disappears for all nouns. At the end of the 11th Century, the case-marking system of English is still intact as a system, in that the same case categories are involved, but the evidence supporting the category distinctions are now greatly reduced because of widespread syncretism of forms. In Peterborough, located at the southern border of the Northeast Midlands area, it appears that the case-marking system at the end of the first third of the 12th Century is not radically different from OE in the category distinctions which it makes, but syncretism of forms have brought the system very close to extinction. Although the distinction between dative and accusative still maintains a tenuous hold, the distinction is no longer marked in the feminine or plural pronouns, or in the nouns, where the old dative inflection has now become reanalyzed as a post-prepositional inflection. Verbal selection of genitive objects either have disappeared entirely or are at least unusual in this area. The Final Continuation12 shows that by the middle of the century, the dative/accusative distinction 11. Final nasals were already lost in some morphological contexts, such as in the infinitive, in the early Northumbrian texts. 12. The second lengthy addition added to The Anglo-Saxon Chronicle, dealing with years 1132-1154. The first lengthy addition, dealing with years 1122-1131, is called the First Continuation. 8 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/7 DOI: https://doi.org/10.25810/k43m-zz07 Loss and Consequence 9 have entirely disappeared, and the distinction between subjects and objects is no longer marked in the determiner system, (12) through (14). (12) & benam ælc ðone riht hand and deprived each-UN the-ACC/DAT right hand-UN ‘and deprived each of them of their right hands’ (PC 1125.9) (13) him me hit beræfode him man it-ACC/DAT bereaved ‘He was deprived of it’ (PC 1124.51) (14) & iærnde ða þurh him & ðurh ealle his freond and asked then through him and through all his friends namcuþlice þone abbotrice namely the-ACC/DATabbacy-ACC/DAT ‘And then asked, through him and all his friends, specifically for the abbacy.’ (PC 1127.49) In all these examples, the verb formerly required or at least allowed an object in the genitive case; for example, the verb of (13), benimen ‘deprive’, would have earlier normally had a deprive in the accusative or the dative, and an object of deprivation in the genitive, although a minority pattern with a dative deprive and an accusative object of deprivation already exists at the OE stage. These objects are now indistinguishable from ordinary direct objects, appearing either unmarked or in the old accusative form. Not surprisingly, no genitive objects are found in the Final Continuation. Further south, the dative/accusative distinction was still quite healthy even towards the end of the 12th Century, although the category distinction is only optionally marked. Objects are still frequently marked with genitive case. However, there is limited evidence available from the original texts from the southern part of the country to reach firm conclusions about how the loss of case marking proceeds. In view of the fact that only a few remnants of the old case-marking system are to be found in the Final Continuation of the Peterborough Chronicle around the middle of the 12th Century, it is no surprise to find that most of these remnants have disappeared entirely by the beginning of the 13th Century in the northeastern part of the county. The nominal suffix –e is still found fairly frequently in the area’s texts, namely the Ormulum—a long poem written by a monk with the intention of explaining the gospels in English. A distinct dative form is hardly ever used for plural nouns, where a single form has usually been generalized to all cases, and such examples found are restricted to the objects of prepositions. With singular nouns, the most frequent use of the old dative suffix is again on the objects of prepositions, but the inflection is also found in some other positions, namely the complements of nouns and adjectives, (15) and (16): 9 Walters: Loss and Consequence Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 10 (15) unþeaw gode laþest vice god-DAT loathsome-est ‘A vice most loathsome to God’ (Ormulum A f.54) (16) cristeto wurðmund Christ-DAT to glory ‘for the glory of Christ’ (Ormulum M f.25.10) However, such examples are not common, and most non-prepositional uses of the dative are to be analyzed as fixed expressions which can be listed in the lexicon. Most importantly, there are no examples of the –e inflection on what is the equivalent of the indirect object in ModE which are not preceded by a preposition. Numerous examples show that the unmarked form could be used for the ‘indirect’ object as well as the direct, and that the indirect object never had dative inflection (17) and (18): (17) Ha chepeð hire sawle þe chapmon of helle she sells her soul the merchant of hell ‘She sells her soul to the merchant of hell’ (AW 213.28) (18) Ne talde ha þen engel na tale not told she the angel no tale ‘She did not tell the angel any tale’ (AW 35.30) Despite the lack of case marking, the recipient and theme are not yet distinguished by a fixed word order. As has been evidenced, a distinct inflectional category which could be called ‘dative’ still exists, but this case is now an exclusively syntactic case which is not lexically selected by any verb. In the pronominal system, the distinction between the accusative and dative has been completely lost in the feminine, neuter,13 and plural pronouns. The dative forms hire (feminine) and ham (plural) have entirely supplanted the old accusative forms. In the neuter, however, it is the old accusative form which has replaced the dative form. Although examples with a neuter dative are rare, the few examples which are to be found shown that the dative form has already been replaced by the accusative form (19): (19) nis hit neod Zeorde? not-is it-UN need rod? ‘It is (i.e. the child) not in need of a rod?’ (AW 167.23) 13. Gender is by this time normally natural gender, not grammatical, although in this dialect few remnants of the old system in the use of feminine pronouns to refer to some nouns which historically have feminine gender are found. 10 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/7 DOI: https://doi.org/10.25810/k43m-zz07 Loss and Consequence 11 This pronoun would have been in the dative case in this construction in OE. A distinction between the old accusative form of the neuter pronoun, hit, and the old dative form, him, has lasted longer than did the dative/accusative distinction with any other pronoun. If this distinction in forms continues to reflect a category distinction between accusative and dative case, it would be evidence that this distinction persists in the grammar much longer than suggested by previous scholarship [cf. Allen 1995]. 3. The Comparisons to English’s Germanic Cousins The loss of case-marking distinctions in OE is a surprisingly orderly and systematic affair, as it has been attempted to show. The loss of the accusative/dative distinction does not immediately result in the loss of all lexical case markings. Proposed dative experiencers continue to flourish for a long period with the ‘impersonal’ verbs. However, unlike English, many of its Germanic cousins did not suffer from this type of reduction, or at least they did not suffer this reduction type until much later. It should be noted at this time, due to the length of this paper, the comparisons made amongst the Germanic languages will not be representative of a complete grammatical characterization. Mainly, a brief description of nominal and pronominal markings will be discussed. In Gothic, the original nominative singular ending of masculine a-stem nouns in Proto-Germanic was *-az. Of all the Germanic languages, Gothic has remained closest to this, with its suffix –s (20): (20) Goth. OHG dags ‘day’ tag (Walters) The Gothic nominative plural of the same class has the ending –ôs, which significantly differentiates Gothic from some—but not all—other Germanic languages (21): (21) Goth OGH fuglôs ‘birds’ fogala (Walters) The third person singular masculine personal pronoun in Gothic is is. This form differentiates Gothic from a number of languages in which that pronoun begins with h- (22): (22) Goth OE is hē (Walters) 11 Walters: Loss and Consequence Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 12 Unlike some other Germanic languages, Gothic regularly distinguishes between the accusative and dative cases in the first and second person singular pronouns (23): (23) Goth OE ACC mik mē ‘me’ DAT mis mē ‘me’ ACC þuk ðē ‘thee’ DAT þus ðē ‘thee’ (Walters) Contrasted against Gothic, Old Norse (ON) preserves the ending *-az of Proto- Germanic in the nominative singular both of masculine a-stem nouns and of most strong masculine adjectives as –r, by way of runic –ar(24): (24) ON Goth OHG armr arms arm ‘arm’ góðr gôþs guot ‘good’ (Walters) The nominative plural of the same masculine a-stems (although not of the adjectives) is expressed by means of the suffix –ar (25): (25) ON Goth OHG armar armôs arma ‘arms’ fuglar fuglôs fogala ‘birds’ (Walters) In the masculine and feminine third person personal pronouns, ON shows forms beginning in h-, unlike several of the other languages, including Gothic (26): (26) ON Goth hann is ‘he’ honum imma ‘him’ (dat. sg.) hon si ‘she’ hennar izôs ‘her’ (gen.sg.) (Walters) Like Gothic, but unlike a number of the other languages, ON normally shows a distinction between accusative and dative in the first and second person singular personal pronouns (27): (27) ON OE ACC mik mē ‘me’ DAT mér mē ‘me’ 12 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/7 DOI: https://doi.org/10.25810/k43m-zz07 Loss and Consequence 13 ACC þik ðē ‘thee’ DAT þér ðē ‘thee’ (Walters) Old Frisian shows no ending for the nominative singular masculine a-stem nouns, nor for the nominative singular of masculine strong adjectives (28): (28) OF Goth wei wigs ‘way’ gôd gôþs ‘good’ (Walters) The nominative plural ending of the masculine a-stem nouns is variable in OF, alternative between –ar or –er, although sometimes with –a. However, in some of the western dialects, it is -a and –an or –en, rather than –ar or –er. And as in OE, OF has third person personal pronouns beginning with h- throughout (29): (29) OF Goth hi is ‘he’ him imma ‘him’ (dat. sg.) hiu si ‘she’ hire izôs ‘her’ (gen. sg.) hit ita ‘it’ (Walters) Like OF, Old Saxon masculine nominative singular ending of both a-stem nouns and strong adjectives disappears completely (30), contrasting sharply with Gothic: (30) OS Goth OF dag dags dei ‘day’ gôd gôþs gôd ‘good’ (Walters) The nominative plural of the masculine a-stem nouns in OS is –os (31): (31) OS Goth OHG fuglos fuglôs fogala ‘birds’ (Walters) The masculine third person personal pronoun in OS shows forms beginning with h- in the nominative singular, with much less frequent occurrence in such forms in other cases. Although this feature distinguishes OS clearly from Gothic, the Saxon forms are also different from ON, which shows much more widespread use of h- (32): (32) OS Goth ON hê is hann ‘he’ 13 Walters: Loss and Consequence Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 14 imu imma honum ‘him’ (dat. sg.) siu si hon ‘she’ ira izôs hennar ‘her’ (gen. sg.) (Walters) Additionally, most OS texts do not distinguish between accusative and dative in the first and second person singular personal pronouns (33): (33) OS Goth ACC mî mik ‘me’ DAT mî mis ‘me’ ACC thî þuk ‘thee’ DAT thî þus ‘thee’ (Walters) Again, contrasting with Gothic, Old Low Franconian, in the nominative singular of masculine a-stem nouns, shows no ending (34): (34) OLF Goth day dags ‘day’ (Walters) (The same holds true for the masculine nominative singular of strong adjectives.) All undisputed masculine a-stem nominative plurals show the ending –a (35): (35) OLF Goth daga dagôs ‘days’ (Walters) The only personal pronoun in OLF that shows an initial h- is the masculine nominative singular (36): (36) OLF Goth he is ‘he’ hie imma ‘him’ (dat. sg.) (Walters) And as in OS, there is no distinction between accusative and dative in the first and second person singular personal pronouns in OLF (37): (37) OLF OS Goth ACC mi mî mik ‘me’ DAT mi mî mis ‘me’ ACC thi thî þuk ‘thee’ 14 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/7 DOI: https://doi.org/10.25810/k43m-zz07 Loss and Consequence 15 DAT thi thî þuk ‘thee’ (Walters) And finally, Old High German is also contrasted against Gothic because OHG shows no trace of the original *-az ending of the nominative singular in the masculine a-stem nouns (38): (38) OHG Goth tag dags ‘day’ (Walters) OHG has also lost the same ending in the masculine nominative singular of strong adjectives. There a new ending –êr is frequently found, which has been added by analogy to the demonstrative pronouns: blint or blintêr ‘blind’ (Walters). The nominative plural of the masculine a-stem nouns in OHG is regularly –a (39): (39) OHG Goth berga bergôs ‘mountains’ fugala fuglôs ‘birds’ (Walters) In the third person personal pronouns, OHG in general diverges sharply from the other Germanic languages by having no forms in h- (40): (40) OHG ON OF ër hann hi ‘he’ sīn honum him ‘him’ (dat. sg.) siu hon hiu ‘she’ ira hennar hire ‘her’ (gen. sg.) iz hinn hit ‘it’ (Walters) OHG harshly diverges once again when dealing with the accusative and dative of the first and second person singular personal pronouns (41): (41) OHG OE OLF OS Goth ACC mih mē mi mî mik ‘me’ DAT mir mē mi mî mik ‘me’ ACC dih ðē thi thî þuk ‘thee’ DAT dir ðē thi thî þuk ‘thee’ (Walters) 15 Walters: Loss and Consequence Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 16 4. The Consequences of Reduction and Why English Changed Despite a common belief to the contrary, the loss of case-marking distinctions was a surprisingly orderly and systematic affair, at least in the south of England where substantial records from the period document the disappearing distinctions. Although many treatments of case marking in ME have given the impression of widespread confusion, the picture which emerges from a systematic study of individual texts is one particular form encroaching on the territory of other forms while category distinctions remain pretty much intact despite the existence of widespread syncretism. These findings go against at least certain variants of the hypothesis that the simplifications of morphology which occurred in ME were due to creolization; for example, they are inconsistent with the idea that deterioration of the case-marking system of English was mainly due to incomplete language-learning on the part of French speaker who failed to master the case-marking system of English. A number of writers on OE have suggested that inflections became largely non- functional as a result of the growing Anglo-Norse contact: as Anglo-Norse contact grew, the case-marking system in OE atrophied and were lost. From this loss, the concomitant development of fixed word-order resulted. Bradley’s (1904) discussion of the issue is old and informal, but still seems to be highly lucid and full of good sense: Let it be imagined that an island inhabited by people speaking a highly inflected language receives a large accession of foreigners to its population. […] In our imaginary island, the foreigners will soon pick up a stock of words; if the island language is like the Germanic ones, in which the main stress is never on the inflexional [sic] syllables, their task will be much easier. The grammatical endings will be learnt more slowly, and only the most striking will be learnt at all. The natives will soon manage to understand the broken jargon of the new comers, and to adopt it in conversation with them, avoiding the use of those inflexions which they discover to be puzzling to their hearers. But if they acquire the habit of using a simplified grammar in their dealings with foreigners, they will not entirely escape using it in their intercourse with each other. If there is intermarriage and absorption of the strangers in the native population, the language of the island must in a few generations be deprived of a considerable number of inflexional [sic] forms. Let us now consider a somewhat different case. Suppose that the two peoples who live together and blend into one, instead of speaking widely distinct languages, speak dialects not too far apart to allow of a good deal of mutual understanding from the first, or at any rate as soon as the ear has been accustomed to the constant differences of pronunciation. The two dialects, let us suppose, have a large common vocabulary, with marked differences in inflexion—a very frequent case, because phonetic change is apt to cause greater divergences in the unstressed endings than in the stressed stems of words. The result will be much the same as when peoples speaking distinct languages are mingled; indeed there are reasons for thinking that the change will be even more rapid and decisive. For one thing, the blending of the two peoples is likely to 16 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/7 DOI: https://doi.org/10.25810/k43m-zz07 Loss and Consequence 17 take place more quickly. Then, as the speakers of neither dialect will be disposed to take the other as their model of correct speech, two different sets of inflexional [sic] forms will for a time be current in the same district, and there will arise a hesitation and uncertainty about the grammatical endings that will tend to render them indistinct in pronunciation, and hence not with preserving. (26-28) However, it should be noted that Anglo-Norse contact did not trigger the developments in OE, but merely augmented or accelerated existing tendencies, themselves largely a consequence of the Germanic fixing of stress on the first syllable. With each invasion, OE changed more, similar to the hypothetical accounting given by Bradley (1904: 26-28). And due to these changes, OE systematically evolved into a language that did not need a case-marking system to differentiate its constituents. These lost inflections resulted in a more stabilized syntax.14 Although a detailed discussion comparing OE and its Germanic cousins could not be included (dates of change in particular), due to length, it should be obvious that OE changed significantly while its cousins did not necessarily change, at least not at the same rate that OE changed. In order to do such a detailed study, much more research—as well as length—would be required. It has been hypothesized that OE changed due to the number of invasions that occurred. With each invasion came a difference in language, resulting in changes of the ‘native’ language. And as the invaders’ languages might not have been of Germanic origin, the systems present within Proto-Germanic quite possibly could not have been maintained. 14. For further discussion on the development of a stable word order, see Bean 1983. 17 Walters: Loss and Consequence Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 18 Abbreviations ACC = accusative DAT = dative EME = Early Middle English GEN = genitive Goth = Gothic INSTR = instrumentive MASC = masculine ME = Middle English ModE = Modern English NOM = nominative OE = Old English OF = Old Frisian OHG = Old High German OLF = Old Low Franconian OS = Old Saxon PL = plural SG = singular UN = unmarked Ælc.P = Homilies of Ælfric: A Supplementary Collection. (see References) Cited by homily and line number. Ælc.Th = The Homilies of the Anglo-Saxon Church. (see References) Cited by volume, page and line number. AW = The English Text of the Ancrene Riwle: Ancrene Wisse. (see References) Cited by page and line number. PC = The Peterborough Chronicle 1070-1154. (see References) 18 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/7 DOI: https://doi.org/10.25810/k43m-zz07 Loss and Consequence 19 Appendix A Table 3. OE General Masculine Declension Sg. Pl. Nom. se cyning ‘the king’ þā cyningas Acc. þone cyning þā cyningas Gen. þæs cyninges þāra cyninga Dat. Instr þæm, þÿ cyninge þæm cyningum Table 4. OE General Neutral Declension Sg. Pl. Nom. Acc. þæt scip ‘the ship’ þā scipu Gen. þæs scipes þāra scipa Dat. Instr þæm, þÿ scipe þæm scipum Table 5. OE General Female Declension Sg. Pl. 1. Nom. sēō talu ‘the tale’ þā tala Acc. þā tale þā tala Gen. þære tale þāra tala Dat. Instr. þære tale þæm talum 2. Nom. sēō glōf ‘the glove’ þā glōfa Acc. þā glōfe þā glōfa Gen. þære glōfe þāra glōfa Dat. Instr. þære glōfe þæm glōfum Table 6. OE The –an Declension Masc. Fem. Neut. Sg.Nom. se guma ‘the man’ sēō byrne ‘the coat of mail’ þæt ēāge ‘the eye’ Acc. þone guman þā byrnan þæt ēāge Gen. þæs guman þære byrnan þæs ēāgan Dat. Instr þæm, þÿ guman þære byrnan þæm, þÿ ēāgan Pl.Nom.Acc. þā guman þā byrnan þā ēāgan Gen. þāra gumena þāra byrnena þāra ēāgena Dat. Instr. þæm gumum þæm byrnum þæm ēāgum Note: There are going to be exceptions, as with every language, but in the interest of length, the general paradigms are the only ones discussed and/or illustrated 19 Walters: Loss and Consequence Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 20 Appendix B Table 7. Declension of adjectives in OE Masculine Feminine Neuter A.‘Strong’ adjectives* Example: til ‘good’ Singular—all genders NOM til tilu til ACC tilne tile til GEN tiles tilre tiles DAT tilum tilre tilum Plural—all genders NOM tile tila tilu ACC tile tila tilu GEN tilra tilra tilra DAT tilum tilum tilum B. ‘Weak adjectives** NOM tila tile tile ACC tilan tilan tile GEN tilan tilan tilan DAT tilan tilan tilan Plural—all genders NOM tilan ACC tilan GEN tilra, -ena DAT tilum *If stem is long, the –u feminine nominative singular and the neuter nominative/accusative plural is deleted, as in gōd ‘good.’ **Very roughly, the ‘strong’ form of an adjective was used when the adjective was not preceded by a determiner, and the ‘weak’ form was used when a determiner was present. 20 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/7 DOI: https://doi.org/10.25810/k43m-zz07 Loss and Consequence 21 Table 8. Paradigm of the definite determiner in OE. Masculine Feminine Neuter Singular—all genders NOM se sēo þæt ACC þone þā þæt GEN þæs þæ”re þæs DAT þæ”m þæ”re þæm Plural—all genders NOM þā ACC þā GEN þāra DAT þæ”m 21 Walters: Loss and Consequence Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 22 References ALLEN, C. 1995. Case Marking and Reanalysis: Grammatical Relations from Old to Early Modern English. Oxford: Clarendon Press. ALLEN, C. 1986a. “Dummy Subjects and the Verb-Second ‘Target’ in Old English.” English Studies 6: 465-470. ANDREWS, A. 1985. “The Major Functions of the Noun Phrase.” In Timothy Shopen (ed.) Language Typology and Syntactic Description, 62-154. Cambridge: Cambridge University Press. BEAN, M. 1983. The Development of Word Order Patterns in Old English. London: Croom Helm. BRADLEY, H. 1904. The Making of English. London: Macmillan. CAMPBELL, A. 1959. Old English Grammar. Oxford: Clarendon Press. GNEUSS, H. 1996. Language and History in Early England. Brookfield, VT. HAIMAN, J. 1974. Targets and Syntactic Change. The Hague: Mouton. HEALEY, A. AND R. VENEZKY. 1980. A Microfiche Concordance to Old English. Toronto: Centre for Medieval Studies, University of Toronto. HELFENSTEIN, J. 1870. A Comparative Grammar of the Teutonic Languages. London: MacMillan and Co. HOPPER, P. AND S. THOMPSON. 1980. “Transitivity in Grammar and Discourse.” Language 56: 251-299. KOOPMAN, H. AND D. SPORTICHE. 1990. “The Position of Subjects.” Lingua 85: 211- 258. LIGHTFOOT, D. (ed.). 2002. Syntactic Effects of Morphological Change. Oxford: Oxford University Press. MILROY, J. 1997. “Internal vs. External Motivations for Linguistic Change.” Multilingua 16: 311-323. MOORE, S. 1928. “Earliest Morphological Changes in Middle English.” Language 4: 238-266. 22 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/7 DOI: https://doi.org/10.25810/k43m-zz07 Colorado Research in Linguistics 6-2004 Loss and Consequence: An Examination of the Old English Case Marking System As Opposed to that of Other Old Germanic Languages Denise E. Walters Recommended Citation Against Optional wh-Movement