Britta Irslinger Intensifiers and reflexives in SAE, Insular Celtic and English Abstract: Intensifiers and reflexives have been studied as features both in areal linguistics and in the context of substratum hypotheses. While typical SAE languages differentiate between intensifiers and reflexives, English, Welsh and Irish use complex intensifiers for both functions. This article discusses the two strategies with regard to their diachronic developments, starting with PIE. Complex intensifiers are first recorded in Old British and emerge only later in English and Irish. These complex intensifiers are then increasingly used as reflexives, constituting an instance of areal divergence from SAE between the late Middle Ages and the early modern period. Breton, on the other hand, maintains its intensifier – reflexive differentiation due to areal convergence. Keywords: intensifier, reflexive, diachronic developments, origins, linguistic convergence, areal linguistics; Insular Celtic, English, SAE Britta Irslinger: Albert-Ludwigs-Universität Freiburg; britta.irslinger@mail.uni-freiburg.de 1 Introduction European linguistics, and in particular its areal linguistics studies, have dedicated much attention to intensifiers and reflexives. The presence of such a distinction was indicated as a membership criterion for the SAE-Sprachbund (Haspelmath 2001), and its absence was attributed to Brythonic influence for English1 and to a prehistoric Afro-Asiatic substratum for the North-Western languages (Vennemann 2013). All these hypotheses are based on the claim that the similarities between certain languages or language groups are due to linguistic contact. Complex intensifiers , their origin and their functions, are crucial to all these hypotheses. They consist of a pronoun plus a second element. The pronoun agrees with its antecedent with regard to person, number and gender, and, within 1 See e. g. Tristram 1999: 24; Vezzosi 2005: 228ff.; Vennemann 2013: 122. Poppe (2009: 253ff.), who also takes Irish into account, concludes that the hypothesis of British influence on Old English remains unproven, albeit attractive. On the other hand, Lange (2005: 269f.; 2007: 186) criticizes Vezzosi’s theory. 10.1515/if-2014-0010 160 Britta Irslinger the older languages, it also inflects for case. The second element is either a simple intensifier, a reflexive adjective, or a noun, which may or may not occur alone. The present paper discusses the expression of reflexivity typical for SAE languages from a diachronic perspective, starting with Proto-Indo-European. ’oncentrating on French and German as representative of SAE, we will consider the main developments which lead to the systems found in the modern languages. Features which did not survive until the modern period will also be treated. The paper then examines the rise of complex intensifiers in ‘rittonic, Irish and English and their subsequent employment as reflexives. All developments will be dated as precisely as possible to establish a time-frame against which the propositions of the different hypotheses can be checked. 1.1 Welsh, Irish and English vs. SAE Typical SAE-languages differentiate between reflexives and intensifiers. Their distribution is shown on the following map from Haspelmath 2001: 1501. Fig. 1. Languages with intensifier-reflexive differentiation The differences between the two types are illustrated with examples from German and English. German expresses reflexivity with the pronouns mich, dich, sich, etc. (ex. 1a, 1b), while uninflected selbst self is used as an intensifier (ex. 1c adnominal, 1d adverbial), English uses my-, your-, himself, etc. in both cases.2 2 See König 2001: 747ff.; König & Siemund 2000: 40ff.; Haspelmath 2001: 1501. Intensifiers and reflexives in SAE, Insular Celtic and English (1) 161 a. Ich sehe mich im Spiegel. I see myself in the mirror. German English b. Er spricht ständig mit sich. He keeps talking to himself. German English c. Der Präsident selbst wird mit uns sprechen. The president himself will talk to us. German English d. Der Präsident wird sich um die Sache selbst kümmern. The president will himself deal with the matter. German English Reflexives and intensifiers in Modern Welsh and Modern Irish are similar to the English ones. Welsh hun or hunan is preceded by a possessive pronoun, Ir. féin by an independent or object personal pronoun. The pronouns are inflected for person, number and gender3 (cf. Table 1). Table 1. Paradigms of reflexives/intensifiers in Modern Welsh and Irish Modern Welsh North South sg pl 1 2 3m 3f 1 2 3 fy hun dy hun ei hun ei hun ein hun eich hun eu hun Modern Irish fy hunan dy hunan ei hunan ei hunan ein hunain eich hunain eu hunain mé féin tú féin é féin í féin muid/sinn féin sibh féin iad féin The following examples from Modern Welsh (ex. 2a–2d) and Modern Irish (ex. 3a–3d) illustrate the uses of these forms as reflexives (ex. 2a, 2b, 3a, 3b) and as adnominal (ex. 2c, 3c) and adverbial intensifiers (ex. 2d, 3d).4 (2) a. Gwelaf fy hun yn y drych. see.ϿЁЅYSЅό ϿϱЅЅYSЅό ϔϩА in the mirror I see myself in the mirror. (reflexive) 3 Welsh uses identical intensifiers for the 3rd person singular, i. e. there is no gender distinction (King 2003: 96). 4 Examples (2a–3d) from Poppe 2009: 254, who also discusses alternative expressions. See Thomas 1996: 268; King 2003: 96f.; ‘orsley, Tallerman & Willis 2007: 220f. for Welsh, and NIG: 86; GG: 122ff., 131f.; Nolan 2012: 70ff. for Irish. See Ó ’urnáin 2007: 2, 1305ff. for examples from Iorras Aithneach (’onnemara Irish). 162 Britta Irslinger b. Mae e’n siarad is â’i hun. he-ϿА talking.Цϩ with-ϿϱЅЅYkЅό ϔϩА He keeps talking to himself. (reflexive) c. Bydd yr arlywydd ei hun yn siarad â ni. be.υЛАYkЅό the president ϿϱЅЅYkЅό ϔϩА ϿА talking.Цϩ with us The president himself will talk to us. (adnominal) d. Bydd yr arlywydd yn delio â’r mater hwn ei hun. be.υЛАYkЅό the president ϿА dealing.Цϩ with-the matter this ϿϱЅЅYkЅό ϔϩА The president will himself deal with this matter. (adverbial) (3) a. Feicim mé fhéin sa see.ϿЁЅYSЅό ϿЁϱϩYSЅό ϔϩА scathán. in-the mirror I see myself in the mirror. (reflexive) b. Tá sé ag caint leis féin. be.ϿЁЅ ϿЁϱϩYkЅόYϨ at talking.Цϩ with.kЅόYϨ ϔϩА He is talking to himself. (reflexive) c. Labhróidh an t-uachtarán é talk.υЛА féin linn. ϿЁϱϩYkЅόYϨ ϔϩА with.kϿϣ the president The president himself will talk to us. (adnominal) d. Pléifidh an t-uachtarán é deal.υЛА the president féin leis an gceist. ϿЁϱϩYkЅόYϨ ϔϩА with the matter The president will himself deal with the matter. (adverbial) When following a possessive adjective and a noun, Welsh ei hun means own (ex. 4a). Irish féin is used similarly, but without the personal pronoun (ex. 4b). (4) a. Eich car eich hun ’dy hwn, ϿϱЅЅYlϿϣ car ϿϱЅЅYlϿϣ ϔϩА is ’te? this.ЅόYϨ then Is this your own car, then? b. Chóirigh Mairéad a (King 2003: 98) gruaige féin. brush.ϿЁА Mairéad ϿϱЅЅYkЅόYυ hair Mairéad brushed her own hair. ϔϩА (Nolan 2012: 85) An important difference between English, Welsh and Irish is the separability of the constituents. In English, the pronouns and -self are fused into one word. In Welsh, although both elements are represented as individual words, hun never occurs without a possessive adjective. In Irish, the presence or absence of the pronoun distinguishes in some cases the reflexive use of féin (ex. 5a) from the intensifier (ex. 5b). However, reflexive féin occurs without a pronoun after a conjugated preposition (ex. 5c and 3b above) or a Intensifiers and reflexives in SAE, Insular Celtic and English 163 possessive adjective + verbal noun (ex. 5d, 5e). In the latter example, á originates from the fusion of do to and the 3rd pers. sing. masc. possessive adjective. (5) a. Mholfainn mé féin praise.ΫϱϩβYSЅό ϿЁϱϩYSЅό ϔϩА I would praise myself b. Mholfainn (Nolan 2012: 76) féin é. praise.ΫϱϩβYSЅό ϔϩА ϿЁϱϩYkЅόYϨ I myself would praise him c. Chuir mé (Nolan 2012: 77) ceisteanna orm put.ϿЁА ϿЁϱϩYSЅό questions féin. on.SЅό ϔϩА I put questions to myself (lit. I put questions on myself. ) (Nolan 2012: 81) d. Tá mé do mo thuirsiu féin. be.ϿЁЅ ϿЁϱϩYSЅό to ϿϱЅЅYSЅό tiring.Цϩ ϔϩА. I am tiring myself e. Bhí sé (Ó Dónaill 1977: 533, s. v. féin) á bhearradh féin. be.ϿЁА ϿЁϱϩYkЅόYϨ to-ϿϱЅЅYkЅόYϨ shaving.Цϩ ϔϩА He was shaving himself. (NIG: 86) The intensifier féin normally occurs without a pronoun whenever the person is already indicated otherwise. This is the case after a non-emphatic personal pronoun (ex. 5f) or a possessive adjective + noun (ex. 4b above). Simple féin is further used as a coreference marker after a conjugated preposition (ex. 5g and 3b above) or as an intensifier after a conjugated verb form (ex. 5h). (5) f. Cé agaibh is sine tú féin nó Síle? [ΫϱϿ older]ΫϱϨϿ ϿЁϱϩYlЅό ϔϩА or Síle who of.lϿϣ Who (of you) is older, you (yourself) or Síle? g. Tá go leor aici be.ϿЁЅ a lot féin. at.kЅόYυ ϔϩА She has a lot herself. h. Déan (GG: 123) (GG: 124) féin é. do.ϔϨϿYlЅό ϔϩА it Do it yourself. (GG: 122) Féin can also be used as a focus particle with the meaning even, only (ex. 5i, 5k).5 Adding the pronoun avoids ambiguity as it makes clear that féin is to be understood as an intensifier (ex. 5j, 5l). 5 See König 2001: 747 on focus particles. 164 (5) Britta Irslinger i. an t-easpag féin the bishop even even the bishop (NIG: 87) j. an t-easpag é the bishop féin ϿЁϱϩ.3ЅόYϨ ϔϩА the bishop himself (NIG: 87) k. dá bhfeicfinn féin iad if see.ΫϱϩβYSЅό ϔϩА them even if I saw (had seen) them l. dá bhfeicfinn mé (NIG: 87) féin iad if see.ΫϱϩβYSЅό ϿЁϱϩYSЅό ϔϩА them if I saw them myself (NIG: 87) The intensifier féin with a disambiguating pronoun is regularly used after a noun (ex. 3c, 3d above, but note the order of constituents in cleft sentences, ex. 5m) or after an emphatic personal pronoun (ex. 5n).6 Such cases show the doubling of the person, which can also be found in English and Welsh. (5) m. Ba é an t-easpag féin a was ϿЁϱϩYkЅόYϨ the bishop bheannaigh an scoil. ϔϩА ЁζϣYϿА blessed the school It was the bishop himself who blessed the school. n. Chonaic mise saw mé (NIG: 86) féin é. ϿЁϱϩYSЅόYζϨϿϑ ϿЁϱϩYSЅό ϔϩА it I myself saw it. (GG: 124) 1.2 Substratum influence in Brittonic? In his most recent article, Vennemann (2013: 134) explains the lack of the intensifier-reflexive differentiation in Insular ’eltic and English as the result of substratal influence. He claims that the feature passed from Afro-Asiatic into ’eltic and from there into English. He also claims that the same development has taken place with regard to the North Sea Germanic (Ingvaeonic) languages spoken on the continent, as Afro-Asiatic originally spread along the coasts and influenced the ’ontinental ’eltic languages which supplanted it. He contrasts the long-term developments of reflexive expression in Insular ’eltic and English with those of Romance and German, and classifies the individ- 6 See Ó Siadhail 1989: 228. Intensifiers and reflexives in SAE, Insular Celtic and English 165 ual developments in these languages, inclulding PIE, according to the reflexivity typology elaborated by Heine (2005) for Creole languages (Vennemann 2013: 136ff.). He concludes that while Insular Celtic and English have undergone several changes due to the alleged direct and indirect substratum influence, Romance and ”erman continued the PIE type without any modification. However, Vennemann (2013: 137) is quite mistaken with regard to the situation in PIE, which he assigns to Heine’s (2005: 207) type B. Languages of this type use personal pronouns to express reflexivity in the 1st and 2nd persons, but have a special reflexive pronoun in the 3rd person as e. g. “rench or ”erman. The situation in PIE was more complex than a simple dichotomy between reflexive pronouns and other pronominal intensifiers, as PIE possessed also the inflectionally marked category of middle voice. Reflexivity is one of the semantic categories found within the functional field of reflexivity/middle which includes also anticausative, impersonal and passive.7 In Modern ”erman and modern Romance languages, most verbs belonging to these situation types are accompanied by the same marker, i. e. the reflexive pronoun. This was however not the case in Late PIE, where verbs of body care and body motion as well as anticausatives, just to name a few, used the middle voice.8 The reflexive pronoun was used only as a full reflexive , i. e. to mark situations in which the agent is at the same time the initiator and the endpoint of a prototypically other-directed transitive event like He kills himself (Kemmer 1993: 52). The inflectionally marked category of middle voice as inherited from PIE developed in Latin, Celtic and ”othic into the passive voice. On the other hand, in Romance and ”ermanic, the original domain of the middle voice was taken over by the reflexive pronouns. However, the scope and degree of their grammaticalization as middle markers shows some variation within the individual languages.9 7 See Kemmer 1993; Haspelmath 2003: 233ff.; Cotticelli Kurras & Rizza 2013: 8ff. 8 See Tichy 2004: 85; Clackson 2007: 142f.; Meiser 2009: 318ff.; Haspelmath 2003: 235; Cotticelli Kurras & Rizza 2013: 9f. 9 “or the developments from Latin to Romance see Hofmann & Szantyr 1965: 174f., 293ff.; “lobert 1975: 382ff., 386ff.; Miller 1993: 228ff.; Kemmer 1993: 151ff.; Haspelmath 2003: 234f. “or ”ermanic see Kemmer 1993: 182ff. Harbert 2007: 317ff., 322ff.; 166 Britta Irslinger 2 Reflexives from PIE to SAE 2.1 Reflexives and intensifiers in Proto-Indo-European As the individual IE languages do not provide constistent evidence with regard to a pronominal reflexive marker, different hypotheses have been proposed with regard to the reconstruction of the situation in PIE. According to the traditional approach, the evidence points to the reconstruction of two stems for the reflexive pronoun, i. e. *s(e) e- and *se-. The latter can be found, beside *s e-, in reflexive pronouns in Italic, Germanic, ‘altic and Slavic, and also in nominal formations, namely in Greek ἕτα ο companion , based on the stem *se-to-, and perhaps in the Welsh intensifier hun self < *s( )o -no(Schrijver 1997: 83). Many scholars assume that these stems originally had different functions.10 ‘eekes (1983: 215ff.; 2011: 236) claims that *se- was the anaphoric-reflexive pronoun, while the 3rd person of the possessive pronoun is based on the adjective *s o- own . Then, according to ‘eekes, *s o- was an intensifier, which over time became one with the pronoun. Hackstein (2003: 73) assumes instead that the two functions were expressed by means of different case forms of the pronominal stem *s e-: the nominative was used as an intensifier while the other cases were reflexives. However, as the reflexive refers to the subject, most scholars deny the existence of this nominative for logical reasons (e. g. Dunkel 2014: 761). Hackstein argues that the main function of the reflexive pronoun, i. e. the expression of coreference with the subject, does not exclude the existence of the nominative at all. The nominative refers to the subject as an apposition, which is normally also the case for intensifiers. In ModHG er hat es selbst getan, he did it himself , selbst (albeit uninflected) is analyzed as a nominative. Even though this nominative, which is reconstructed as *s e , did not survive in any language, several languages used it to build intensifiers, e. g. Ved. svay-ám -self < *s e óm, Alb. vetë person, self < *s e -tah₂ or OIr. féin -self < *s e sine.11 The validity of these reconstructions is still under debate, but the alleged existence of the nominative *s e is not crucial to the argument. Even if one denies it, the intensifiers can alternatively be reconstructed as being based on other cases of the reflexive pronoun. Thus the accusative could be the base for the OIr. féin < 10 See the survey and discussion in Puddu 2005: 65. 11 Reconstructions by Klingenschmitt 1994: 241f. Intensifiers and reflexives in SAE, Insular Celtic and English 167 *s e-sin (Schrijver 1997: 83ff.), and the dative for the Vedic svay-ám -self < *s o óm for himself , while Albanian vetë person, self < *s e-to- (Dunkel 2014: 759) is a nominal derivation from the stem. Most literature on the subject agrees that the reflexives of the older IE languages do not indicate person, number or gender (which is still the case e. g. in Indo-Iranian, ‘altic, and Slavic languages as well as in Ancient ”reek, at least partially). To avoid ambiguity, it was possible to use the 1st and 2nd person pronouns instead of PIE *s( )e-.12 As a result of this strategy, which was common in many languages, the invariable reflexive pronoun became restricted to the 3rd persons, as e. g. in ”erman and “rench, where even today the pronoun is identical for singular and plural (Hackstein 2003: 63). However, many of these claims were challenged by recent studies. According to Puddu (2005: 88), PIE *se- marked coreference, while *s(e) e- marked possession and was also used as an intensifier. The former developed into a reflexive because of its function as a marker of topic continuity. The latter, which worked as the basis for the development of the possessive pronouns, entered the pronominal declension through the creation of a genitive of possession. It evolved from a possession marker to a marker of coreference with the subject. ’otticelli Kurras & Rizza (2013: 10f.) claim that in Late PIE the verbal strategy of middle voice covered the entire reflexivity/middle spectrum, and that only some reflexive functions received extra pronominal marking. Over time, this latter strategy became more and more prominent until it eventually replaced the middle voice. According to them, the pronominal markers found in the oldest attested IE languages are innovations originating from several different sources. Puddu (2005: 118ff.) argues that pronominal reflexives based on the coreference marker *se- are restricted to the languages located in a contiguous area in central Europe, i. e. ”ermanic, Italic, ‘altic and Slavic. This marker was originally restricted to the 3rd persons as in Latin, ”ermanic and Lithuanian. The pronominal markers based on *s e-, which are found e. g. in Vedic and Avestan, are to be interpreted as possessive markers or intensifiers. Only in Sanscrit, svá- has become a true reflexive pronoun just like in later stages of Avestan. This secondary strategy developed in addition the primary strategy, i. e. the use of reflexive markers with nominal heads. The function of *s e- as an intensifier is also clearly shown by the Vedic form svay-ám -self and by compounds like Vedic svar j- self-ruling and Young Avestan aθβarəšta self-created (Puddu 2005: 69f., 116). 12 See ‘rugmann 1911: 397ff.; Hackstein 2003: 69; Petit 2001: 28. 168 Britta Irslinger On the other hand, ’eltic, where *s e- is attested as an intensifier, did not possess a pronominal reflexive marker. As outlined above, the use of *se- as a reflexive is a rather recent development shared by Germanic, Italic, ‘altic and Slavic. ’eltic, situated at the Western periphery of this linguistic area, did not take part in the innovation. Interestingly, Ingvaeonic lost the Proto-Germanic reflexive marker *sik as a result of its contact with ’eltic (Puddu 2005: 186). Puddu s argument is based on her analysis of Insular ’eltic, which is attested only from the middle of the first millenium AD. The lack of evidence for reflexive markers in ’ontinental ’eltic seems to be due to the fragmentary attestation of the individual languages. It remains thus possible that ’eltic, which is after all adjacent to or even partly overlapping with the *se-area, acquired the reflexive in a first stage but lost it again like Ingvaeonic.13 Later on, Insular ’eltic and English create reflexives from intensifiers and thus undergo the same development as the other IE languages. The creation of reflexives from intensifiers is widespread cross-linguistically (König & Gast 2009: 154f.). ’rucially, the morphology of the intensifiers in Irish and Welsh does not show any non-IE influence. In both languages they contain exclusively well known IE elements, but they do not share a common origin. As mentioned above, Schrijver (1997: 83) claimed that Welsh hun self and its ‘reton and ’ornish cognates (see below, 3.3.1) derived from *s( )o -no-, with initial *s - or *s-. The semantics of this form can be compared to Old ’hurch Slavonic svojь own , and Ved. svayám -self . The suffix *-no-, however, is paralleled by Gothic seins, OHG sīn his < *se -no-.14 Schrijver assumes that, due to its formal similarity, hun was subsequently attracted to the numeral un, unan one , which generated the variant hunan. A competing etymology explains hun as an instance of the numeral un, unan one < *o no-. The h- in hun is either assumed to be unetymological (Fleuriot 1964b: 259; Fleuriot 1964a: 216) or to result from a compound with the reflexive pronoun, i. e. *s e-o no- (Morris Jones 1913: 307). However, neither explanation is fully convincing. Thus, it remains unclear why un one would appear both with and without h-. In addition, there is no known example across languages of a compound consisting of a reflexive and the word for one (Parina 2007: 395). The same may be true for a compound consisting of an intensifier *s eand the numeral one . 13 Note that the map in Puddu 2005: 225 is rather imprecise with regard to the presumable diffusion of ’eltic in prehistoric times. 14 ’f. also Old Icelandic sveinn, ME swein etc. boy, lad < *s o -no-, Lith. sváinis brother-in-law, wife s sister s husband < *s o -ni o-, originally perhaps member of the household/belonging to the household (Vries 1962: 567). Intensifiers and reflexives in SAE, Insular ’eltic and English 169 2.2 Reflexive pronouns in German, Latin and French Reflexivity is expressed by ordinary pronouns in the 1st and 2nd person and by a special reflexive pronoun in the 3rd person, which is identical for singular and plural. Table 2. Reflexive pronouns in Modern High German and Modern French 1sg. 3sg. masc. 3sg. fem. 3pl. ModHG ModFr Ich wasche mich. Er wäscht sich. Sie wäscht sich. Sie waschen sich. Je me lave. Il se lave. Elle se lave. Ils se lavent. I wash myself. He washes himself. She washes herself. They wash themselves . The same situation can already be found in Latin. In the 2nd century AD, however, the 3rd person reflexive pronoun is also used for the other persons in formulaic expressions in legal texts (ex. 6) (Pieroni 2010: 436):15 (6) ego L. Marius I, Ampliatus sibi et suis libertis L. Marius.ϩϱϨ Ampliatus.ϩϱϨ ЁζυϣYβΞА and ϿϱЅЅYβΞАYϿϣ freedman.βΞАYϿϣ libertabus=que posterisque freedwoman.βΞАYϿϣ=and descendants.βΞАYϿϣ I, L. Marius Ampliatus, to me, to my freedman and freedwoman and descendants. (IL’V: 600; 2nd cent. AD) In German as well there are numerous examples of sich used for the 1st and 2nd persons, dating from late MHG until the modern period16 (ex. 7). The phenomenon is frequent in dialects of south-eastern (ex. 8) and central Germany. The origin of this use is controversial. Hackstein (2003: 69) seems to assume that it was inherited from PIE, while Paul (1919: 130f.) and ‘ehagel (1923: 305) explain it as a later transfer of the 3rd person reflexive pronoun, like it happened in Latin.17 15 See Hofmann & Szantyr 1965: 176; ’ennamo 1991: 3ff. Examples of the expansion of se to the 1st and 2nd persons can also be found in Romance dialects (Puddu 2005: 143). 16 See Paul 1919: 130 also with later examples. ’ontrary to Paul, there is no evidence for the combination of the 1st person subject ich and the 3rd person reflexive sich in Goethe s Italienische Reise, chapter 3. The correct reading of the passage in question is: Der Postillon eilte mehr, als ich wünschte: er hatte noch keine Messe gehört und wollte sie in Innsbruck, es war eben Marientag, um desto andächtiger zu sich nehmen. 17 See Puddu 2005: 142 on Yiddish zik, which is used for all persons. This development has been explained as an influence from Slavic languages. 170 Britta Irslinger (7) … gehe ich in IFG. Zimmer, sich go.ϿЁЅYSЅό ϿЁϱϩYSЅό into his princely grace s room IFG. zu unterreden mit ЁζυϣYkЅόgϿϣ with his princely grace to speak … I go into his princely grace s room to speak with his princely grace (Schwein. 1, 268; 1577) (8) Modern ‘avarian Mir sehe sich. we see.ϿЁЅYSϿϣ ЁζυϣYgЁζΫϔϿYkЅόgϿϣ We see ourselves/each other. (reflexive or reciprocal; Hackstein 2003: 69) 2.3 Replacement of the 3rd person reflexive pronoun by personal pronouns ‘oth French and German have, in some cases, replaced the 3rd person reflexive pronoun with personal pronouns. 2.3.1 French In Modern French, soi, the stressed form of the reflexive pronoun se, can refer only to indefinite subjects, while for definite subjects the non-reflexive personal pronouns are used. In Old French, sei, soi can refer to all subjects (ex. 9, 10), but instances of the modern usage can already be found (ex. 11):18 (9) tut sun aver qu’ od sei en ad portet all ϿϱЅЅYkЅό wealth Ёζϣ with ЁζυϣYЅόgϿϣ there bring.ϿЁЅYkЅό all his wealth which he has brought with him (Alexius p. 59, v. 19, l. 91; c. 1050) (10) or l’ at od sei now her have.ϿЁЅYkЅό with ЁζυϣYЅόgϿϣ Now he has her with him (11) Mes son (Alexius p. 67, v. 122, l. 609; c. 1050) lyeon avoec lui but ϿϱЅЅYkЅόYϨ lion with ot ЁζυϣYkЅόYϨ have.ϿЁАYkЅό ‘ut he had his lion with him (Yvain 6518, c. 1177–1179; ms. c. 1235) 18 See Rheinfelder 1976: 103f.; Picoche & Marchello-Nizia 1998: 233f. Intensifiers and reflexives in SAE, Insular Celtic and English 171 2.3.2 German In Old and Middle High German, the reflexive sih is inflected for number and case, as well as for gender in the singular form. Its use is restricted to genitive sg. masc./neut. and accusative sg./pl. In all other instances personal pronouns are used instead, cf. ex. (12) from OHG, ex. (13–15) from MHG.19 Table 3. Paradigms of reflexive pronouns in Old and Middle High German. Personal pronouns used as reflexives are given in italics. masculine sg. pl. (12) N G D A N G D A – sīn imu sih OHG neuter – sīn imu sih – iro im sih Oh fimf dumbo feminine masculine – ira iru sih – sîn im(e) sich MHG neuter feminine – sîn im(e) sich – ir(e) in sich – ir ir(e) sich intfanganēn liohtfazzon ni nāmun oli mit but five foolish.ϩϱϨYϿϣ received.βΞАYϿϣ lamps.βΞАYϿϣ ϩζό take.ϿЁАYkϿϣ oil with in ЁζυϣYβΞАYϿϣ The five foolish ones but did not take any oil with them for the lamps they had received (Tatian 148, 2; c. 830) (13) er nam im manige schouwe an mislîchen buochen he take.ϿЁАYkЅό ЁζυϣYβΞАYkЅόYϨ many look on various he took many a look at various books (himself) books (Heinrich v. 6; c. 1190) 19 See ‘raune & Reiffenstein 2004: 241 for OHG, Paul 2007: 215 and ‘esch 1967: 295f. for MHG and Early ModHG. Ex. (14, 15) are from the ’odex St. Georgen LXIV, which was probably written in Freiburg i. ‘r. (Karlsruhe, ‘ad. Landesbibliothek). For a description see ‘esch 1967: 33ff. 172 (14) Britta Irslinger brot … das allen wolgeluste und suͤssekait in im bread het which all pleasures beschlossen and sweetness in ЁζυϣYkЅόYϩ include.ϔϩυ have.ϿЁАYkЅό bread … which included in itself all pleasures and sweetness (Hs. 33, 2,9; ‘esch 1967: 295; 1383) (15) dar an im as gott selber wolgevallen het by ЁζυϣYkЅόYϨ god ϔϩА pleasure as god himself was pleased by himself have.ϿЁАYkЅό (Hs. 33, 3,4; ‘esch 1967: 295; 1383) Only from later Middle High ”erman onwards (15th c.), the dative pronouns are replaced by the accusative reflexive sich. This usage becomes generalized in the 16th century, and is adopted into the written standard. The older usage survives in regional varieties, especially in the south-west.20 2.3.3 Reflexives and intensifiers based on nouns Vennemann (2013: 125ff.) compares the Welsh and English complex intensifiers and reflexives to those of various Afro-Asiatic languages. These include reflexives and intensifiers which originated from common nouns meaning self , body , person etc. ’laiming that English and Insular ’eltic stand alone within the IndoEuropean language family regarding this section of grammar (ibid.: 124), Vennemann takes Afro-Asiatic as the only possible model on which Welsh hun and successively English self were calqued. He claims that both words were originally nouns and that an even closer parallel, which employs the noun pen head , is found in the Welsh idiom ar ei (dy etc.) ben ei (dy etc.) hun(an) on his (your) own, by himself (by yourself etc.) (King 2003: 97). Literally, this phrase means on his own head . According to ”P’ (2729, col. 2), however, this idiom, which can be found in Modern Welsh, is only attested from 1588 onwards. It is highly questionable that English self and Welsh hun originate from nouns. Lange (2007: 61) argues that there are no occurences of *min self (i. e. possessive + the noun self ) and that plural marking on self is nonexistent before the mid-16th century. These features support the assumption that self was an 20 See Paul 2007: 215; ‘esch 1967: 295f. for MH”; Ebert et al. 1993: 215, § M 64 for Early ModH”; “rings 1957: 141 for regional varieties. Intensifiers and reflexives in SAE, Insular Celtic and English 173 adjective.21 This is also the most probable origin of Welsh hun. The reconstructed form *s( )o -no- contains PIE *‑no-, a suffix typically used to create adjectives (see above 2.1 and fn. 14). Irrespective of this, the grammaticalisation of intensifiers and reflexives from nouns denoting the body or body parts is crosslinguistically widespread.22 Puddu (2005: 90ff., 225) identifies reflexive markers with a nominal head as the primary strategy of the neighbouring Eastern Indo-European languages Indian, Iranian and Tocharian. The nouns in question are Vedic ātmán- breath, soul, self , Vedic tan -, Avestan tanū- body , Tocharian ‘ kektseñ obl. body (Hackstein 2003: 73f.), Tocharian ‘ ṣañ āñm, A ṣñi āñcäm one s own self (Pinault 2013). These nouns may or may not be combined with a possessive pronoun. Further, this type of reflexive marking can co-occur as a secondary strategy in addition to other marking types23 or be restricted to certain persons.24 The most recent and geographically closest parallels are found in European languages, e. g. in Old and Middle High German, Old French, Middle English and ‘asque. 2.3.3.1 Old and Middle High German In ex. (16), from Old High German, it remains open to debate whether the phrase líp minaz my body replaces a reflexive pronoun or if it should rather be understood literally as my person or my body .25 In Middle High German, the noun līp body , in combination with the possessive pronoun or a name in the genitive, could be used as a personal pronoun or a reflexive, cf. ex. (17).26 According to Jakob Grimm & W. Grimm (DW‘: 12, 582, s. v. Leib), the function of this phrase is, at least in some contexts, equivalent to the combination of a reflexive pronoun and the intensifier selbst. 21 The etymology of OE self , OHG selb < Germanic *selba- is controversial. Perhaps it originates from P’l. *sel- o/ā- (what is) owned (?) from the root *selh₁- to take (LIV²: 529), cp. OIr. selb f. ā, Welsh (h)elw m. property, possession (Kluge & Seebold 2011: s. v. selber, citing also an alternative etymology). The PIE suffix *- o/ā- forms mainly adjectives. The use of an adjective meaning possessing as an intensifier is paralleled by German eigen and English own. 22 See Heine & Kuteva 2002: 57ff. s. v. ‘ODY, 168 s. v. HEAD. 23 See ’otticelli Kurras & Rizza 2013: 9; Puddu 2005: 113 with examples from Hittite. See Puddu 2005: 133 with an example from Homeric Greek. 24 See Puddu 2005: 103f. with examples from Armenian and Albanian. 25 The passage is based on Luke 23,42: [memento mei domine] cum veneris in regnum tuum. 26 Jacob Grimm (1866: 265f.) claims that Old French ses cors is a calque on the German phrase. 174 (16) Britta Irslinger joh laz I thas líp minaz in sconi ríchi thinaz body ϿϱЅЅYόζϩYSЅό in beautiful realm ϿϱЅЅYόζϩYlЅό surrender.ϿЁЅYSЅό the I surrender myself / my person / my body to your beautiful realm (Otfrid IV 31,20; 868 AD) (17) eʒ bekumberte mīnen līp it afflict.ϿЁАYkЅό ϿϱЅЅYΞΫΫYSЅό body.ΞΫΫ it afflicted myself (Iwein 345; c. 1200) 2.3.3.2 Old French In Old French, the noun cors body preceded by a possessive pronoun, can be used instead of a personal pronoun, a reflexive pronoun (ex. 18, 19) or an intensifier (ex. 20). The expression refers to the person indicated by the possessive pronoun. It is a stylistic device to convey polite or occasionally derisive distance. This use disappears in Middle French, and its relics can be found today in Modern French in the fixed expression à son corps deféndant reluctantly, unwillingly . The nouns char flesh and persone person are employed in the same contexts, albeit less frequently.27 (18) Autretant l’ so much aim come mon him love.ϿЁЅYSЅό like cors ϿϱЅЅYΞΫΫYSЅό body I love him like myself (19) cunduit sun (Yvain 3792; 1177-1179) cors en la presse des Francs take.ϿЁЅYkЅό ϿϱЅЅYΞΫΫYkЅό body in the crowd of-the Francs he takes himself into the crowd of the Francs (20) joster i vait ses (Rol. 3370; 1075- 1110) cors mèísmes fight.ϔϩυ here go.ϿЁЅYkЅό ϿϱЅЅYϩϱϨYkЅό body ϔϩА he goes himself to fight (Isembart [Voretzsch 15, 144]); end of 11th c.) 2.3.3.3 Middle English The Old French locutions with cors discussed in the previous section were imitated in Late Middle English. ’ombinations of possessive pronoun + body, person and self are attested from the late 13th to the late 15th century. Expressions like his own body or his own self are emphatic equivalents of himself (ex. 21). Occasion- 27 See Stéfanini 1962: 331ff.; Rheinfelder 1976: 162; De Lage 1990: 90; Picoche & Marchello-Nizia 1998: 234. Intensifiers and reflexives in SAE, Insular ’eltic and English 175 ally, his own body represents he himself, as do the expressions by ’haucer (ex. 22) and Shakespeare (ex. 23; Mustanoja 1960148f.): (21) where he his oghne body lay where he ϿϱЅЅYkЅόYϨ own body lay where he himself lay (22) my joly (”ower, ’A ii 3468; 1386–1392) body schal a tale telle ϿϱЅЅYSЅό beautiful body will I myself will tell a tale (23) as thy a tale tell (’haucer, ’T, ML 1185; early 1390ies) sweet selfe grow’st as ϿϱЅЅYlЅό sweet self grow.ϿЁЅYlЅό as your sweet self you grow (Shakespeare, Sonnet cxxvi 4; 1609) 2.3.3.4 Basque With its lack of true reflexive pronouns, ‘asque constitutes an interesting case. Here reflexive noun phrases are formed by combining the possessive, i. e. the genitive of an intensive personal pronoun, with buru head and the so-called definite article ‑a (Table 4 and ex. 24, 25) (Trask 2003: 159; Rijk 2008: 113f., 364ff.). Table 4. ‘asque reflexives 1S 2S intimate 2S unmarked 3S 1P 2P 3P (24) neure neure burua heure burua zeure burua bere burua geure burua zeren burua zere(n) burua myself yourself yourself himself, herself ourselves yourselves themselves burua ϨЭYϔϩАYόζϩ head.βζυ myself (lit. the head of mine ) (25) Ispiluan ikusi dugu geure mirror.ϣϱΫ see ΞЛЬ burua our.ϔϩАYόζϩ head.βζυ We saw ourselves in the mirror The intensive personal pronouns themselves are constructed from the ordinary personal pronouns (absolutive or genitive) plus demonstratives. Like demonstra- 176 ‘ritta Irslinger tives, intensive personal pronouns are divided into proximal, (h)au(r) this , medial (h)ori that (just there) and distal (h)ura that, yonder (Trask 2003: 123). The intensifiers for the 1st and 2nd persons thus have a threefold differentiation. However, those for the 3rd persons do not, and vary from region to region, the most widely used being bera, composed of ber- plus the article (Trask 2003: 152). Table 5. ‘asque intensifiers 1S 2S intimate 2S unmarked 3S 1P 2P 3P (26) proximal medial distal neu heu zeu bera geu zeuek bera/berak nerau herori zerori nihaur hihaur zuhaur gerok zerok guhaur zuihauk Neuk I myself you yourself you yourself he himself, she herself we ourselves you yourselves they themselves asmatu dut hori ϔϩАYSЅόYζЁόϿЁϱЬ think ΞЛЬ that I thought that myself (27) Igone bera Igone herself On the one hand, ‘asque parallels Old French, Middle High German and Middle English in using a noun denoting a body-part preceded by a pronoun within the reflexive noun phrase. On the other hand, it is also similar to Irish, as its intensive personal pronouns are formed with the help of demonstratives. Irish féin < *s esin contains the anaphoric particle sin (see below 3.2.2). The latter occurs also as a constituent of demonstrative constructions (GOI: 301ff.). 2.3.3.5 Brabants Dutch ‘rabants Dutch has a reflexive which originated from a complex intensifier (ex. 28). This intensifier is a combination of a possessive pronoun and the attributive possessive intensifier eigen own (König 2001: 755). As Parina (2007: 395) points out, in this respect ‘rabants Dutch is typologically close to Welsh. (28) Ik wash m’n eigen. I wash my own I wash myself. (König 2001: 755) Intensifiers and reflexives in SAE, Insular Celtic and English 177 3 Intensifiers and reflexives in North Sea Germanic and Insular Celtic Several West-Germanic languages have lost the 3rd person reflexive pronoun and replaced it with the ordinary personal pronoun. These languages (Old English, Old Frisian, Old Saxon and Old Low Franconian) are also known as North Sea Germanic or Ingvaeonic. 3.1 English In the following examples from Old English (König & Siemund 2000: 45), whether the 3rd person pronoun has a reflexive meaning or not (ex. 29, 30) can only be deduced from the context. To indicate coreference unambiguously, the pronouns can be combined with the intensifier self /seolf /sylf (ex. 31). (29) hine he bewerað mid wæpnum ϿЁϱϩYΞΫΫYkЅόYϨ ϿЁϱϩYϩϱϨYkЅόYϨ defendYϿЁАYkЅό with weapons he defended himself with weapons (Ælfric, Gr. 96.11–2; late 10th –early 11th c.) (30) ða behydde Adam hine & his wif eac swa dyde and hide.ϿЁАYkЅό Adam ϿЁϱϩYΞΫΫYkЅόYϨ and ϿϱЅЅYkЅόYϨ wife also so do.ϿЁАYkЅό and Adam hid himself and his wife did the same (Ælfric, Gen. 3.9; late 10th –early 11th c.) (31) Hannibal … hine Hannibal selfne mid atre acwealde. ϿЁϱϩYΞΫΫYkЅόYϨ self.ΞΫΫYЅόYϨ with poison kill.ϿЁАYkЅό Hannibal killed himself with poison. (Orosius IV.11; late 9th c.) ’onstituted by a pronoun + self /selves, the Modern English complex intensifiers display a mixed paradigm. In fact, the pronoun component in the 1st and 2nd persons is identical with the possessive pronouns (myself, yourself, ourselves, yourselves), while the 3rd persons are built using the dative forms of the personal pronoun (himself , herself, itself, themselves). In Old and Middle English, alternative forms can be found, including e. g. nominative he self, accusative hine selfne or genitive his selfes. Dative pronouns + self are the most frequent type in Middle English about 1250 (Lange 2007: 61). The dative 1st and 2nd person pronouns are usually superseded by the respective possessives by the end of the 13th century, 178 Britta Irslinger although the type me self can still be found as late as in the 15th and 16th centuries (Mustanoja 1960: 146).28 König & Siemund (2000: 72) describe the following development for Old English, referring to dating taken from Keenan 1996. Sequences of dative pronoun + nominative self came to be interpreted as semantic units around 1000 AD. Later on, they become a phonological word and, eventually, a syntactic unit. The process was complete around 1275. These combinations of pronoun + intensifier adopted the distribution and the meaning of both their components, i. e. they came to be used as intensifiers as well as reflexives. According to van Gelderen (2000: 52ff., 249) complex intensifiers are used as reflexives for 3rd persons earlier than for 1st and 2nd persons. Further, the increase of complex intensifiers used as reflexives depends on the dialect. They are more frequently found in southern texts (which are also later). Lange (2007: 152) states that complex intensifiers first appear in Early Middle English (1150) in two environments: either as adjuncts, forming a constituent with a subject, or as arguments in subject position. In the latter case, they occur mostly in a formulaic phrase, e. g. as (he) himself said , restricted to a specific text type. Lange (2007: 162f.) disagrees with Keenan s (1996: 17) claim that the simple intensifier self disappeared during the 1200s. She states that, although Keenan s claim that the attestations of unbound self decrease drastically is correct, examples can still be found in mid-14th century texts. The expression of reflexivity is not affected by the development of complex intensifiers (Lange 2007: 152). The compound reflexive x-self is attested from 1150 onwards, but it still takes a long time until its grammaticalization is complete and it becomes the only reflexive marker in use. Although the instances of compound reflexives outnumber the simple ones already by the end of the 15th century, they completely replace the plain pronouns as reflexive markers only in the 17th century (Lange 2007: 173f.; Peitsara 1997: 288). 28 The dative forms of the pronouns of the 1st and 2nd pers. sg. me and þe developed into mi and þi within a phonological process that also affected other words (e. g. be to bi in biforen, bitwene). As the dependent possessives min and þin lost their -n in the course of Middle English, the new dative forms mi and þi were reanalysed as possessives (Mustanoja 1960: 146, 164). On the syntactic aspects of the replacement see Lange 2007: 77ff. On the inflection of self see Lutz 2002. Intensifiers and reflexives in SAE, Insular Celtic and English 179 3.2 Other West Germanic languages Unlike English, the other Germanic languages of the group in question have reconstituted the intensifier-reflexive differentiation. Dutch borrowed new third person reflexives (zich, zichzelf ) in the 15th century from Middle Low German, which in turn got it from High German (Postma 2012: 140f; Harbert 2007: 179). Frisian and Afrikaans differentiate with the help of complex markers. Frisian uses the simple personal pronoun as a reflexive marker (ex. 32), but can also combine it with the intensifier sels self (ex. 33). The compounds are emphatic and contrastive, but they are always reflexive and cannot be used as ordinary intensifiers (ex. 34) (Harbert 2007: 180). (32) se wasket har she washes herself (or her). (33) se wasket harsels she washes herself (34) ik seach de minister sels I saw the minister himself In Afrikaans, the complex marker expresses reflexivity (ex. 35), while self alone is an intensifier (ex. 36) (König & Siemund 2000: 67). (35) Elke ma moet haarself die volgende afvraa. Every mother should ask herself the following. (36) Sy het die rokkie self gemaak. She made the dress herself. 3.3 Old Irish 3.3.1 Reflexives ‘ecause it is better documented, it will be helpful to consider Old Irish first. As in Old English, reflexivity is expressed through the ordinary personal pronouns. In Old Irish, when a personal pronoun is used with a verb it is always unstressed (except as predicative nominative). These unstressed pronouns are generally either reduced to single phonemes, which are attached to a pretonic preverb, or infixed between preverb and verb.29 Reflexivity can be expressed unambiguously for the 1st and 2nd persons, cf. ex. (37, 38), but not for the 3rd persons. Whether a 3rd person infixed pronoun is reflexive (ex. 39) or not (ex. 40) can only be deduced from the context. 29 See GOI: 255, further 256 on the position of infixed pronouns, and 257 on the forms of infixed pronouns, which are subdivided into three classes. 180 (37) Britta Irslinger no-m·ísligur ϿА-ϿЁϱϩYΫϣΞЅЅYΞYSЅό·abase.ϿЁЅYSЅό (Wb. 17d22; 8th c.) I abase myself (38) fo-t·chridigther=su ϿЁζЦ-ϿЁϱϩYΫϣΞЅЅYΞYlЅό·gird.ϿЁЅYlЅό=lЅό you gird yourself (39) (Ml. 101c3; c. 800–850 AD) ar-nda·cumcabat for-ϿЁϱϩYΫϣΞЅЅYΫYkϿϣ·raise.ЅЛΪϟYkϿϣ in order that they may raise themselves (Ml. 46a12; c. 800–850 AD) (40) co-ndid·moladar so that-ϿЁϱϩYΫϣΞЅЅYΫYkЅόYϨ·praise.ϿЁЅYkЅό (Wb. 16d1, 8th c.) so that he praises him 3.3.2 Intensifiers OId Irish has several intensifiers, which are inflected for gender and number but not for case, i. e. féin, fadéin, céin, cadéin (GOI: 306f.). The forms with initial fcontain PIE *s e- as their first element.30 Table 6. Etymologies of Old Irish intensifiers 1/2S 3Sm 3Sn 3Sf 1P 2P 3P féin fesin féin físin fesin fesin fesin < < < < < < < *s *s *s *s *s *s *s e sin e es sin e e(d) sin e sī(s) sin ens sin ens sin and/or *s e is sin? ens sin and/or *s e ens sin fadéin fadesin < < *s e de sin *s e de es sin fanisin fadisin fadesin < < < < *s *s *s *s e de sī(s) sin e nis sin e de is sin e de ens sin As shown by Schrijver s (1997: 74ff.) reconstructions (Table 6), all forms use the anaphoric particle sin the aforementioned as their closing element. With the exception of the 1st and 2nd sg., they also include infixed personal pronouns. Stems starting with fad- additionally contain the particle de. Schrijver (1997: 146) points out that this particle “seems to convey the approximate meaning reference to an 30 See Schrijver 1997: 78 on ca-, which probably originates from *ke even he/she and may be related to OIr. cía, ci- though etc. Intensifiers and reflexives in SAE, Insular Celtic and English 181 item in the preceding context but not to the immediately preceding item; yonder. ” The forms with *-de- seem to have a special contrastive connotation in Wb. and Ml. However, as fadesin is gradually ousted by fesin in Sg., no clear-cut semantic distinction can be found in OId Irish (Schrijver 1997: 79, 81). There are numerous examples attesting the use of these intensifiers. In the Milan glosses, fadéin translates Latin ipse (ex. 41). Ex. (42, 43) illustrate féin etc. used in apposition to the subject expressed in the verbal ending: (41) fadéin gl. ipse (42) in tain ro-n-icub when (Ml. 23c1; c. 800-850 AD) féin ϿЁζЦ-ϩΞЅYЁζϣ-come.υЛАYSЅό ϔϩА (Wb. 18b14; 8th c.) when I myself shall come (43) tuicci feissin understand.ϿЁЅYkЅό ϔϩАYkЅόYϨ (Wb. 12c16; 8th c.) he himself understands Further, féin etc. can optionally be added to disambiguate the reflexive meaning of phrases with infixed personal pronouns, as shown in ex. (44, 45). However, in OId Irish, these intensifiers are not used as reflexive markers for the 1st or 2nd persons yet. This usage only appears in Middle Irish (Dottin 1987: 223), although not yet consistently. In ex. (46), fén (orthographic variant of féin) is an intensifier of the 2nd person reflexive pronoun tú. In contrast to this, in ex. (47), from the same text, the reflexive meaning of fair on himself has to be deduced from the context. In this example, fé(i)n is used instead of fesin for the 3sg.m. During the Middle Irish period, féin is used increasingly for all persons, genders and numbers, and eventually survives alone in the modern language (see eDIL: s. v. fadéin). (44) no-d·moladar fesin ϿА-ϿЁϱϩYΫϣΞЅЅY+YkЅόYϨ·praise.ϿЁЅYkЅό ϔϩАYkЅόYϨ (Wb. 17b 21; 8th c.) who praises himself (45) ní-s·tuarascbat feisin cen gutai ϩζό-ϿЁϱϩYΫϣΞЅЅYYkϿϣ·express.ϿЁЅYkϿϣ ϔϩАYkϿϣ without vowels they do not express themselves without vowels (46) in tan tucais fén tú isin croich when give.ϿЁЅYlЅό ϔϩАYlЅό you in-the cross when you gave yourself in the cross (47) (Sg. 7a 11; 9th c.) ruc fen breith (PH 172; Middle Irish) fair give.ϿЁАYkЅό ϔϩА judgement on-kЅόYϨ he himself has passed judgment on himself (PH 603; Middle Irish) 182 Britta Irslinger The examples above clearly show that the Modern Irish combination of pronoun + féin is still rare in Old and Middle Irish, as object pronouns are infixed and subject pronouns are not obligatory. Only in ex. (48) the subject pronoun is mandatory, because there is no inflected verb. ‘y the 17th century at the latest, Irish reflexives and intensifiers have converged into one, (ex. 49, 50), with the intensifiers é féin and mé féin used as reflexives. (48) mo ṡīur & me féin do breith i mbrait ϿϱЅЅYSЅό sister and ϿЁϱϩYSЅό ϔϩА to carry off.Цϩ into captivity my sister and myself to carry off into captivity (Togail Troí, LL 32085; late 10th c.)31 (49) agus go bhfothruigeadh é and that bathe.ЅЛΪϟYkЅό féin as an anbhruith. ϿЁϱϩYkЅόYϨ ϔϩА in the broth and that he would bathe himself in the broth (Foras Feasa I, 22.8; c. 1618–1634) (50) agus do fholuigh mé mé féin and ϿА hide.ϿЁА ϿЁϱϩYSЅόYЅϟ ϿЁϱϩYSЅό ϔϩА and I hid myself (Genesis 3.10; ‘edell 1685) Due to the lack of research, it is impossible to give a more detailed account of this development and to indicate e. g. at what period complex reflexives begin to outnumber infixed reflexives. 3.4 Brittonic 3.4.1 Intensifiers The Modern Welsh inventory of identical reflexives and intensifiers as presented above (section 1) is only attested from Middle Welsh onwards, i. e. from the late 12th century AD. ‘ack then, however, their distribution was not yet the same as in the Modern period. As Vezzosi (2005: 236f., 238) and Parina (2007: 391ff.) point out, e hun occurs in Middle Welsh predominantly as an intensifier and only occasionally as a marker of coreference.32 In the text Pedeir Keinc y Mabinogi The four branches of the Mabinogi , Parina found 41 instances of possessive + hun al- 31 See ‘reatnach 1994: 268. 32 See also the examples in D. S. Evans 1964: 89f. Intensifiers and reflexives in SAE, Insular Celtic and English 183 together, but only 11 of these occur in reflexive contexts.33 Parina thus concludes, that the original function of possessive + hun is intensification. This finding is corroborated by the corresponding combinations in the other ‘rittonic languages, i. e. Middle ’ornish possessive + honan (ex. 51) (see Lewis & Zimmer 1990: 37) and Middle ‘reton possessive + (h)unan (ex. 52) (see HMS‘: 86f.): (51) ow colon ow honan ϿϱЅЅYSЅό heart ϿϱЅЅYSЅό ϔϩА my own heart (52) ma (RD 2042; 1350–1400) hunan ez stoeyf ϿϱЅЅYSЅό ϔϩА ϿА bow.υЛАYSЅό I myself will bow (”wenolé: 829; 1580) In the earlier period, which was not considered by Parina, possessive + hun is attested three times, in Old Welsh (ex. 53, 54) as well as in Old South-West-‘ritish (ex. 55). In the first two examples, the Latin equivalents confirm that they are used as intensifiers (A. “alileyev 2000: 88): Old Welsh (53) dit-ti-hun to-lЅό-ϿϱЅЅYlЅό-ϔϩА gl. tibi soli (M’: 9 a.a. A. “alileyev 2000: 88; 9th c.) to you alone (54) mi mi-hun ϿЁϱϩYSЅό ϿϱЅЅYSЅό-ϔϩА gl. ipsa I myself/ich selbst (M’: 51 b.a. A. “alileyev 2000: 88; 9th c.) Old South-West-‘ritish (55) dedi hi hun to-kЅόYυ ϿϱЅЅYkЅόYυ ϔϩА to her herself (Angers 477, fo. 74b, Patrol. X’ col. 488; end of 9th c.)34 33 Note that adnominal intensifiers in phrases like the president himself will speak to us (above, section 1.1, ex. 1c) are also called emphatic reflexives (see e. g. Haspelmath 2003: 235). The 11 examples in PKM mentioned by Parina all belong to this type. There are, however, no instances of full reflexives , i. e. possessive + hun does not mark coreference in two-participant events like He kills himself (see above, 1.2 and see Kemmer 1993: 243). 34 See Schrijver 2011: 51; “leuriot 1964b: 259; “leuriot 1964a: 72f. See “leuriot 1964a: 8–11, 27–31 for the description of the ms. 184 Britta Irslinger Examples (53–55) further show that the possessive pronouns and hun self are tightly bound already in the oldest period, which makes them complex intensifiers. Unlike in Irish, the pronoun of the intensifier is not dropped when another subject pronoun (ex. 53, 54) or conjugated preposition precedes it (ex. 55). This pleonastic expression remains unchanged also in later periods, cf. ex. (56) from Middle Welsh: (56) ny adwn ni drwc arnam ny ϩζό let.ϿЁЅYSϿϣ ϿЁϱϩYSϿϣ evil hunein on-SϿϣ ϿЁϱϩYSϿϣ ϔϩАYϿϣ we don t let evil on us ourselves (PKM 21.04–6; 1375–1425) In Middle Welsh, the combination of possessive + hun is occasionally used as a coreference marker with other-directed predicates, cf. ex. (57, 58) (Vezzosi 2005: 238; see also Parina 2007: 390f.; D. S. Evans 1964: 89): (57) ony ledy du hun unless kill.ϿЁЅYlЅό ϿϱЅЅYlЅό ϔϩА unless thou dost kill yourself (58) na chapla dy (Gwyrthyeu, ‘‘’S x 24.35; 1250–1300) hun ϩζό reprove.ϔϨϿYlЅό ϿϱЅЅYlЅό ϔϩА do not reprove yourself (’yngorau, ‘‘’S ii 23.28; 1375–1425) When these predicates are formed with the preverb MW ym-, M’o. ym-, em-, om-, M‘r., Mod‘r. em, which expresses reflexivity or reciprocity (see below, 3.4.2), the combination of possessive + hun/honan/(h)unan is added as a reinforcement or for disambiguation. Middle Welsh (see D. S. Evans 1964: 89) (59) nac ym-hoffa vyth dy hvn ϩζό Ёζυϣ-praise.ϔϨϿYlЅό ever ϿϱЅЅYlЅό ϔϩА do not ever praise thyself (’atwn, ‘‘’S ii 29.37; 1275–1325) Middle ’ornish (see Lewis & Zimmer 1990: 37) (60) y honan ny yl ϿϱЅЅYkЅόYϨ ϔϩА ym-sawye ϩζό can.ϿЁЅYkЅό Ёζυϣ-save.Цϩ he himself cannot save himself (61) rak hacre mernans certan eys em-lathe y for cruel death den my a grys man I (PD 2877–8; 1350–1400) honan ny gaffe certainly than Ёζυϣ-kill.Цϩ ϿϱЅЅYkЅόYϨ ϔϩА ϩζό find.ϔϨϿYkЅό ϿА think.ϿЁЅYϔϨϿζЁЅ For a more cruel death, certainly, than to kill himself no man may find, I think (RD 2073; 1350–1400) Intensifiers and reflexives in SAE, Insular Celtic and English 185 Modern ‘reton (see HMS‘: 272) (62) o veza … m’ en ϿА be.ϔϩυ em brisent re ho-unan as ϿЁϱϩYkЅόYϨ Ёζυϣ evaluate.ϿЁЅYkϿϣ too much ϔϩАYkϿϣ as they have too high an opinion of themselves (JKS: 336; c. 1862) Vezzosi (2005: 239), who argues for ‘rittonic influence on English, claims that a parallel between them regards the creation of the word oneself . The term emerged late in Middle English through the analogical extension of the self - paradigm. An equivalent word, an-unan oneself is present in the Middle period of the ‘rittonic languages. Like oneself, it expresses coreference to an indefinite subject, cf. ex. (63) from Modern ‘reton: (63) lod a zifenne en em lakâd an unan da varner some ϿА forbid.ϿЁАYϔϨϿζЁЅ ϿЁϱϩYkЅόYϨ Ёζυϣ put.ϔϩυ oneself as judge.ϩϱЛϩ some (of those laws) forbade (one) to make oneself a judge (IAI: 122; 1893) While Vezzosi s analysis of the function of an-unan is correct, her attribution of the word to Insular ’eltic and her dating of the example are not. An-unan, the combination of the definite article an and unan, is attested from Modern ‘reton onwards. The language of the example taken from Hemon (HMS‘: 87, note 2) is not Middle ‘reton as claimed by Vezzosi, but Modern ‘reton dating from 1893.35 The creation of ‘reton an-unan might have been influenced by the “rench intensifier soi-même, which is used for indefinite subjects. In Middle Welsh, yr un, the combination of the article and the numeral one , denotes someone, anyone (ex. 64), but not oneself (D. S. Evans 1964: 87).36 (64) A chyt dywettit uot porthawr ar lys Arthur, and though say.ϔϨϿυYϔϿЅ be.Цϩ porter yr vn. nyt oed at court of Arthur.όζϩ ϩζό be.ϿЁАYkЅό βζυΞЁА one And though it was said that there was a porter at Arthur s court, there was none (”eraint, WM 223.17–8; c. 1350) 35 See “avereau 1997: 114f. on an-unan in ‘reton dialects. 36 Note that Vezzosi (2005: 238) adheres to the hypothesis that Welsh hun etc. is identical with the numeral un one . According to the alternative etymology proposed by Schrijver (see above, 2.2.3), hun does not originate from the numeral and Welsh yr un is irrelevant to the present context. ’ontrary to ”P’ (s. v. hwn), hun is etymologically distinct from the demonstrative pronoun hwn(n), which can be used with the article: yr hwn the one (see Schrijver 1997: 29). 186 Britta Irslinger 3.4.2 Reflexivity In all ‘rittonic languages, reflexivity can be expressed by compounding a verb with the preverb OW im-, MW, ModW ym-, am-, OSW‘r. im-, em-, M’o. ym-, em-, om- and M‘r., Mod‘r. em. Together with OIr. imb‑, imm-, this preverb originates from Proto-’eltic *ambi- around, about . The etymology shows that the expression of reflexivity is not the basic function of this preverb, but it results rather from a couple of semantic developments. Vendryes (1927) described these developments with regard to Middle Welsh, but they are also valid for Middle ‘reton and ’ornish. Two semantic groups stand out in Middle and Modern Welsh, namely ym- formations expressing reciprocity (ex. 65–66) and reflexivity (ex. 67–68). (65) a. cyrchu to attack, go to, fetch b. ymgyrchu to attack, fight (one another), gather together (66) a. cael to get, obtain b. ymgael to come together, assemble, meet; to have sexual intercourse (67) a. golchi to wash b. ymolchi to wash oneself ୸ golchi hunan (68) a. cuddio to hide b. ymguddio to hide oneself ୸ cuddio hunan As Vendryes (1927: 57) points out, the reflexive meaning developed from the reciprocal one. The latter, which implies the participation of two antagonists, was reinterpreted as the performance by the subject of a reciprocal, and successively a reflexive, activity. 3.4.2.1 Old South-West British Ex. (58–61) in the previous section illustrate the strategy to turn a non-reflexive verb into a reflexive one by combining it with the preverb MW ym-, am-, M’o. ym-, em-, om-, M‘r. em. Three examples of this type are attested in Old South-West ‘ritish, where the preverb is im(-), em, all of which gloss Latin verbs with the reflexive pronoun se (ex. 69–71) (Fleuriot 1964b: 331f.). (69) im-guparton Ёζυϣ-move away.ϿЁЅYkϿϣ gl. se abdicant they move (themselves) away (Orléans 221; mid of 9th c.) Intensifiers and reflexives in SAE, Insular Celtic and English (70) heuan em doguor bravely Ёζυϣ move forward.ϿЁЅYkЅόYΫϱϩϟ gl. inopportunius se ingerit bravely he moves (himself) forward (71) di im dam-guas to Ёζυϣ ϿЁζЦ-submit.ϿЁЅYkЅόYΫϱϩϟ 187 (‘N lat. 12021; 9th or 10th c.) gl. se iuramento constrixerit he submits himself to (?) 37 (Orléans 221, fo. 114, gl. 181; mid of 9th c.) 3.4.2.2 Breton Middle ‘reton displays a way to express reflexivity that is similar to what was seen above, but involves some modifications. Em is no longer a preverb, but the obligatory reflexive particle . Although em alone indicates reflexivity, especially when used with the participle (ex. 72), it is usually preceded by a dependent or object personal pronoun, agreeing in person, number and gender with the subject (ex. 73) (Le Roux 1957: 254f.). (72) da servicha Doe … ez-of to serve.ϔϩυ ”od em laquet ϿА-be.ϿЁЅYSЅό Ёζυϣ put.ϿϿ I started to serve ”od (lit. I am put myself to serve ”od ) (N 243; end of 15th c.) (73) ouzouch … me m-em to-2Ͽϣ I clem ϿЁϱϩYSЅό-Ёζυϣ complain.ϿЁЅYϔϨϿζЁЅ to you … I complain (N 344; end of 15th c.) In Modern ‘reton the inflected personal pronoun falls in disuse. Reflexivity is either indicated by em alone or by en em, i. e. em preceded by the generalized 3rd person sg. masc. dependent pronoun en.38 As the complex intensifiers are still in use, ‘reton still has the differentiation between reflexives (ex. 74, 75) and intensifiers (ex. 76) which is typical for SAE-languages.39 (74) ’n em zifennit! ϿЁϱϩYkЅόYϨ Ёζυϣ ϣζϩ.defend.ϔϨϿYlϿϣ Defend yourselves! 37 The translation remains uncertain, as the semantic modification by the preverb dam- is not clear (“leuriot 1964a: 141). 38 See Le Roux 1957: 252–7 and 256 for dialectal variants. 39 Haspelmath s (2001: 1501) exclusion of ‘reton from SAE is thus incorrect. 188 Britta Irslinger (75) Ni en em wel. we ϿЁϱϩYkЅόYϨ Ёζυϣ/ЁζΫϔϿ ϣζϩ.see.ϿЁЅYϔϨϿζЁЅ We see ourselves. (reflexive or reciprocal) (76) Va-unan am eus graet. myself.ϔϩАζϩЅ have.ϿЁЅYSЅό do.ϿϿ I have done (it) myself. 3.4.2.3 Cornish Middle ’ornish predominantly uses verbs compounded with the preverb ym-, em‑, om-, which indicate coreference. As they do not contain any indication of what person they refer to, the coreferenced constituent has to be inferred from the context. There are no obligatory pronouns or intensifiers (ex. 77, 78), in fact, the addition of intensifiers (above, 3.4.1, ex. 60, 61) can be considered as rather exceptional, since no more than a handful of examples are attested altogether (see also ex. 79 below). (77) Guetyogh om-previ manly take care.ϔϨϿYlϿϣ Ёζυϣ-prove.Цϩ manly Take care to prove yourselves manly (78) Tav, gays (‘M 1194; 1504) thym the om-brene. be silent.ϔϨϿYlЅό leave.ϔϨϿYlЅό to-SЅό to Ёζυϣ-redeem.Цϩ ‘e silent, leave me to redeem myself (‘M 1252; 1504) Prefixed reflexive verbs cease to be productive in Late ’ornish. The prefix omstill appears as a petrified morpheme in certain verbs like ombla fighting and umthan conceiving (Wmffre 1998: 33f.). The drama The Creacion of the World, written in 1611, contains the older construction with prefixed reflexive verbs (ex. 79), but also an alternative, newer one (ex. 80), where coreference is indicated by the possessive adjective. In ex. (81), reflexivity is expressed by the complex intensifier like in Welsh: (79) rag om-sawya ow to honyn keffrys ow Ёζυϣ-save.Цϩ ϿϱЅЅYSЅό ϔϩА also gwreak ha-w flehys ϿϱЅЅYSЅό wife to save myself as well as my wife and children (80) Gwarnys of (’W: 2374–5; 1611) gans Dew an Tase tha wythell an lesster ma rag warn.ϔϿυYϔϿЅ be.ϿЁЅYSЅό by ow and-my children god the father to sawya ha-w flehys build.Цϩ the vessel βζϨ in order to ϿϱЅЅYSЅό save.Цϩ and-my children I have been warned by God the Father to build this vessel to save myself and my children (’W: 2309–11; 1611) Intensifiers and reflexives in SAE, Insular Celtic and English (81) mothosta maab Deew if-you-are son towle tha 189 honnen doare [of god]όζϩ cast.ϔϨϿYlЅό ϿϱЅЅYlЅό ϔϩА down if thou be the Son of God, cast thyself down (Matthew 4.6; William Rowe, c. 1670–1700) Thus the same development occurred in both Welsh and ’ornish, but in the latter language the process took place considerably later. Here, intensifiers and reflexives only came to be identical by the 17th century. 3.4.2.4 Welsh In Middle Welsh, verbs composed with ym- are also attested (ex. 82), but ym- is not obligatory to express reflexivity, cf ex. (83) with an uncompounded verb (D. S. Evans 1964: 89). (82) nac ym-hoffa vyth dy do not ever praise thyself (83) hvn ϩζό Ёζυϣ-praise.ϔϨϿYlЅό ever ϿϱЅЅYlЅό ϔϩА ony ledy du (’atwn, ‘‘’S ii 29.37; 1275–1325) hun unless kill.ϿЁЅYlЅό ϿϱЅЅYlЅό ϔϩА unless thou dost kill yourself (Gwyrthyeu, ‘‘’S x 24.35; 1250–1300) While in ex. (82) dy hun is used to add emphasis to a prefixed reflexive verb, in ex. (83) it functions as a reflexive. The relative frequency and distribution of the two strategies in Middle and Modern Welsh is yet unknown. The same is true for the ratio of ym- reflexives occurring with and without intensifiers. Until a quantitative examination of reflexive expression in Welsh is carried out, the description of the development in Welsh can only be confined to the following remarks. GP’ (s. v. hun (b)) cites a selection of examples dating from the 13th to the th 18 century. Reflexive ym- verbs continue to be in use until Modern Welsh. Nevertheless, they appear to be replaced by constructions consisting of possessive adjective + verbal noun indicating reflexivity from the middle of the 16th century onwards (ex. 84, 85). Thus ein canmol ein hunein in ex. (85) replaces what was expressed by ymganmol in the 1567 and 1588 editions of the ‘ible (ex. 86). Like in Late ’ornish, ym- is replaced by a possessive adjective (at least before a verbal noun). It is difficult to determine the function of ein hunein in ex. (84) and (85). It could be labelled as an intensifier, while the possessive adejctives preceding the verbal nouns could indicate reflexivity. However, as there is no intensifier in ex. (86), ein hunain could alternatively be a redundant reflexive marker. 190 Britta Irslinger (84) ydd ym ϿА ni yn en twyllau en-hunein be.ϿЁЅYSϿϣ ϿЁϱϩYSϿϣ ϿА ϿϱЅЅYSϿϣ deceive.Цϩ ϿϱЅЅYSϿϣ-ϔϩАYϿϣ we are deceiving ourselves (85) Canys nid ydym for yn ein (KL: vi a, l. 15; 1551) canmol ein hunein drachefn ϩζό be.ϿЁЅYSϿϣ ϿА ϿϱЅЅYSϿϣ commend.Цϩ ϿϱЅЅYSϿϣ ϔϩАYϿϣ again wrth-ych unto-lϿϣ for we commend not ourselves again unto you (86) Canys nid ydym for yn ym-ganmol (2. ’or. 5.12; 1620) trachefn wrth-ych ϩζό be.ϿЁЅYSϿϣ ϿА Ёζυϣ-commend.Цϩ again unto-lϿϣ for we commend not ourselves again unto you (2. ’or. 5.12; William Morgan 1588) The possessive adjective preceding the verbal noun has disappeared in Modern Welsh. See ex. (87) vs. ex. (88) containing identical constructions. (87) ym-ogelwch Ёζυϣ-beware.ϔϨϿYlϿϣ … rag eich condemnio eich beware … of condemning yourselves (88) Dw i hunein from ϿϱЅЅYlϿϣ condemn.Цϩ ϿϱЅЅYlϿϣ ϔϩАYϿϣ am rwystro chi (Taith ’ 38; 1726) rhag niweidio ’ch hunan be.ϿЁЅYSЅό ϿЁϱϩYSЅό about prevent.Цϩ ϿЁϱϩYlϿϣ from hurting.Цϩ ϿϱЅЅYlϿϣ ϔϩА I want to try to prevent you (from) hurting yourself (King 2003: 293) While the previous examples illustrate the development from Middle to Modern Welsh, the question of how reflexivity was expressed in Old Welsh and Old ‘ritish remains unanswered. With regard to ex. (83) above, D. S. Evans (1964: 89) states the following: The original construction would have been ony’th ledy du hun, with du hun supplementing the infixed pronoun ’th, which would be the object of the verb. Evans thus claims that in Welsh, and thus also ‘rittonic, reflexivity was expressed exactly like in Old Irish, i. e. by means of infixed personal pronouns. As in Old Irish, then, the 3rd person pronouns are ambiguous, so that the addition of an intensifier became necessary. After the infixed pronouns were lost, which happened in ‘rittonic earlier than in Irish, the intensifiers adopted the additional function of reflexive markers. While this course of development seems plausible, two questions have to be asked: Is there any evidence for this assumption? And what is the role of the prefix ym- with regard to the expression of reflexivity? At least in Middle Welsh, it does not seem to be an unambiguous indicator of reflexivity. Intensifiers and reflexives in SAE, Insular Celtic and English 191 There are two possible examples attested in Old Welsh. Ex. (89) contains immi-, the OW form of ym- followed by an infixed 3rd personal pronoun and the verb. Schrijver (2011: 49, 61) translates the phrase variably as he besmeared himself and as he besmeared him/her . The lack of context makes it impossible to determine which of the two was the intended meaning of the sentence. The Latin verb underlying the OW gloss is not reflexive and does not occur with an object pronoun either. (89) immi-s-line ϿЁζЦgЁζυϣ]-ϿЁϱϩYkЅόgϿϣ-besmear.ϔϨϿυYkЅό gl. allinebat he besmeared himself/he besmeared him/her (M’, cf. A. Falileyev 2000: 91; 9th c.) Also the phrase immi-t-cel contained in ex. (90) has been interpreted in different ways. (90) ni-choilam immi-t-cel ir nimer bichan gutan ϩζό-believe.ϿЁЅYSЅό ϿЁζЦgЁζυϣ]-ϿЁϱϩYkЅόYϨ-hide.ϿЁЅYkЅό the number small ir maur nimer the big under number I do not believe (that) the small number hides it(self) under the big number (MP; 817 AD) Schrijver (2011: 76; 1997: 155) analyses immi-t-cel as a compound verb *immi-cel-. He reconstructs this verb phrase as *immi-t-en-cel < *ambi-(e)t-en keletī.40 This phrase thus originally contained a particle -t- followed by an infixed 3rd sg. masc. pronoun -en-. The pronoun has disappeared, but due to its original presence the particle -t- was preserved. Otherwise -t- would have been lost before a verb beginning with a consonant. Schrijver (2011: 49) gives an alternative analysis of -t- as an infixed 3rd sg. masc. pronoun -t /d/. This pronoun did not survive into Middle Welsh, which only uses -s- as the 3rd person pronoun. Gender and number of this infixed pronoun remain ambiguous. ‘oth suggestions are plausible in some ways and not in others, which cannot be discussed here. ‘oth of them assume the presence of an infixed pronoun, 40 The particle -t /d/ < Proto-‘r. *t could originate from the PIE connector *eti and, according to Wackernagel s Law, was placed in the second position of the sentence. See also Schrijver 1997: 155, 173; Schumacher 2004: 96, 111. According to Mc’one (2006: 239), however, there is no need to reconstruct an otherwise unsupported preform *ambi-(e)t(i)-en-. *æmbi -e(n)- with the infixed 3rd sg. masc. pronoun *e(n) would have developed to æmbið-e(n)- > *ɪmmɪð. In the latter form, /ð/ was replaced by -t /d/, which occurs after various preverbal particles. This trivial replacement was possible because infixing a pronoun after a preverb was an obsolescent pattern in OW. 192 Britta Irslinger which can have non-reflexive or reflexive meaning, i. e. immi-t-cel hides it(self) . However, the context seems to support the reflexive interpretation. De Mensuris et Ponderibus is a Latin treatise on weights and measures which contains Latin and OW glosses. The Welsh sentence appears in a Latin gloss in which the author argues that a certain amount of smaller measuring units fails to tally with the bigger unit, i. e. the sum of the smaller units does not correspond precisely to the bigger one. Thus, Welsh hides itself would be used metaphorically to express the concept of being contained . According to Schrijver s analysis, the OW expression of reflexivity would then be the same as the Old Irish one, while the preverb immi- would be irrelevant with regard to this feature. In contrast to this, the traditionally accepted interpretation does not assume the presence of an infixed (but vanished) pronoun. Lambert (2003b: 132) segments the verbal complex into imm-it-cel, with the preverb imm- expressing reflexivity followed by the subordinating particle it, yt, i. e. that … hides itself … .41 However, Schumacher s (Schumacher forthc. § 7.1) detailed study42 of this particle, which has another etymology than *-t- in Schrijver s analysis, does not corroborate the assumption that it can stand between a verb and its preverb. Schrijver s analysis contains another weak point: if it is true that reflexivity was already expressed by the preverb immi-, then using the infixed pronoun would not be necessary or even possible. Therefore, it will be useful, to have a closer look at the semantics of OW immi-, which became MW, ModW ym-, am-. In Welsh, ym-, am- became a prefix to form compounds from nouns, adjectives and verbs.43 The original meaning around, about can be found in ymdaith(io) to march (< * to travel around ୸ teithio to travel, journey ). In some cases, the difference between the prefixed forms and their bases is less obvious, e. g. amgyffret to comprehend (୸ cyffret to comprehend, understand, comprise; embrace ), ceisio to seek, ask, try ୸ ymgeisio to try, apply, see . Vendryes (1927: 55) claims that the ym-formations may express greater intensity or insistence. Further, as already mentioned, there are ym-formations that express reciprocity and reflexivity. Vendryes (1927: 57) points to MW ymgelu to hide oneself ୸ celu to hide as an example of reflexive semantics. 41 Already Williams (1931: 240) had claimed that OW immi-t-cel represents *ym-yd-gel and thus corresponds to MW yd ym-gel hides himself . See also A. I. Falileyev 2008: 84. 42 I thank Prof Stefan Schumacher (Vienna) for kindly providing me with the prepublication draft of his forthcoming article. 43 See Vendryes 1927: 49ff. for MW; Thomas 1996: 640; Pilch 1996: 34ff.; Zimmer 2000: 228, 231 for ModW. According to Thomas (loc. cit.), reflexive ym- is productive, while reciprocal ym- has become unproductive. Intensifiers and reflexives in SAE, Insular Celtic and English 193 ‘ut is the reflexive meaning to hide onself , as in ex. (91), the only possible meaning of ymgelu? Vendryes (1927: 57) seems to believe so, for Middle Welsh, and the Geiriadur Mawr s. v. lists the verb for the modern language and gives cuddio hunan as a synonym. Despite this, the GP’ does not list any attestation of celaf, celu after the 18th century.44 Further, non-reflexive ymgelu to hide is well documented in Middle Welsh, cf. exx. (92–96).45 (91) Chwi a’m keissyassawch i ϿЁϱϩYlϿϣ ϿА ϿЁϱϩYSЅό seek.ϿϣЛϿυYlϿϣ ym-gelu mwy ragoch val na allaf ϿЁϱϩYSЅό so.that ϩζό can.ϿЁЅYSЅό Ёζυϣ-hide.Цϩ any more before-lϿϣ You have sought me so that I cannot hide myself longer from you. (YSG: 5483–4, ms. 1375–1425) (92) nyt ym-gelaf heb ef, mi yw’r iarll ϩζό hide.ϿЁЅYSЅό said ϿЁϱϩYSЅόYϨ ϿЁϱϩYSЅό am the lord I do not hide it , said he, I am the Lord (Peredur, Peniarth 4, 35r, c. 137, l. 24; 1350) (93) Nyt ym-gelaf ragot arglwyd athro Bown o Hamtwm wyf ϩζό hide.ϿЁЅYSЅό before-lϿϣ lord i. teacher ‘own o Hamtwn be.ϿЁЅYSЅό ϿЁϱϩYSЅό I do not hide it before you lord teacher (that) I am ‘own o Hamtwm (Y‘H, Peniarth 5, p. 220v, c. 887, l. 26; before 1300) (94) Ac nyt oes dim a allo ym-gelu racdunt. and ϩζό be.ϿЁЅYkЅό thing Ёζϣ can.ЅЛΪϟYkЅό hide.Цϩ before-kϿϣ And there is not anything, he could hide before them. (YL, Peniarth 190, p. 152, l.17; 1346) (95) Ef ni allwys ym-gelu oe uot yn y charu ϿЁϱϩYkЅόYϨ ϩζό can.ϿЁАYkЅό hide.Цϩ be.ϿЁАYkЅό be.Цϩ in ϿϱЅЅYkЅόYυ love.Цϩ hi ϿЁϱϩYkЅόYυ He could not conceal being in love with her (Math uab Mathonwy, WM 51.24–5, Peniarth 4, 26r, col. 102 l. 24–5; 1350) 44 GP’ s. v. ymgelaf, ymgelu refers to ym- and celaf, celu. No attestations are available for the compound verb. 45 Examples (91–96) from Rhyddiaith. 194 (96) Britta Irslinger Darparu a oruc dwyn kyrch nos am ben Arthur a’e prepare.Цϩ ϿА do.ϿЁАYkЅό take.Цϩ attack night upon eissoes nyt ym-gelawd hyny rac however ϩζό hide.ϿЁАYkЅό this arthur. lu. Ac Arthur and his host and before Arthur He did prepare to take a night attack upon Arthur and his host. However (he) did not hide this before Arthur. (Brut y Brenhinoedd, NLW Ms. 3035 [Mostyn 116], p. 104v, l.1; 1350–1400) According to Poppe & Reck (2008: 44), the phrase ni chelaf / nyd ymgelaf raggot I will not hide (it) from you in ex. (92) and (93) is a quasi-formulaic expression, which signals a speaker s polite willingness to provide information. It is employed for three semantically matching Anglo-Norman expressions, and, as it does not occur in the same formulaic way in native texts, it is clearly a loan locution. On the other hand, ex. (94–96) express hiding something before someone in the literal sense, and exemplify a non-reflexive use of ymgelu. Determining the function of the preverb ym- in Old and Middle Welsh is overall problematic. Reflexivity is but one of its possible functions, thus all occurrences should be checked individually against their specific context. ’onsequently, considering immi- in OW immi-t-cel as non-reflexive is a viable option46 and reflexivity is expressed by the infixed pronoun as in OId Irish. 3.4.2.5 Summary The expression of reflexivity in the ‘rittonic languages underwent several changes over the centuries. Although it seems probable that Old ‘ritish once used ordinary infixed pronouns like Old Irish, there is no conclusive evidence to confirm this assumption. There are a handful of Old ‘ritish attestations which employ verbs prefixed by imm-. In two cases from Old Welsh, it is unclear whether these verbs also contain an infixed pronoun, and determining whether reflexivity is expressed by these pronouns or rather by the preverb remains difficult. ’oncerning the following development, South-West-‘ritish and Old Welsh diverge. In Middle ’ornish and Middle ‘reton, the preverb strategy was grammaticalised and became the only means to express reflexivity. The complex intensifiers consisting of possessive + honan/(h)unan occur occasionally with reflexive verbs, but their function remains unchanged. In ‘reton, the prefix is grammaticalized as a reflexive particle, which, in Middle ‘reton, is preceded by a possessive adjective. 46 It is of course possible that ymgelu meant hide oneself in OW, and was subsequently used also with non-reflexive meaning. 195 Intensifiers and reflexives in SAE, Insular Celtic and English In Middle Welsh, the two strategies co-occur. From the middle of the 16th century onwards, verbal nouns of ym- verbs appear to be increasingly replaced by constructions including a preverbal possessive adjective, the verbal noun without ym- and a complex intensifier. In present day Welsh the latter has eventually become the principal reflexive marker. Although Cornish is closer to Breton during the middle period, the developments in Late Cornish are similar to those occurring in Welsh. In some cases, the reflexive preverb ym- is replaced by a possessive pronoun, while in others complex intensifiers are used as reflexive markers. 4 Summary and conclusions 4.1 Loss and replacement of reflexives based on PIE *s( )eContrasting hypotheses claim that PIE *s( )e- was either lost in Insular Celtic or it was never present. The same can be said of PIE *s( )e- as a reflexive marker for all persons and numbers. According to the traditional hypothesis, this function was lost for the 1st and 2nd persons in Latin and Germanic, while Puddu (2005) argues that *se- was used only as a 3rd person reflexive marker from the very beginning. The loss of *s( )e- as a 3rd person marker has been generally acknowledged for the North Sea Germanic languages, a partial loss can be observed also in French and German. Consequently, in this respect, there is no sharp divide between languages which have preserved *s( )e- and others which have not (Table 6). Further, the related developments are difficult to reconstruct, as they may apply only to certain varieties or registers of a language or as they may have been reversed by reconstitutions. Table 7. Replacement and spread or reconstitution of reflexives based on PIE *s( )e- reflexivity expressed by personal pronouns 1st and 2nd persons ? 3rd pers. partly 3rd pers. completely spread or reconstitution of *s( )e1st and 2nd persons 3rd person Latin OFr OHG MHG NHG OE IC + + + + + + + + + + + + (+) + 196 Britta Irslinger In Latin, se is used as a reflexive marker for the 1st and 2nd persons in 2nd century legal expressions. This development can also be found in some regional varieties of Romance languages. In Old “rench, the stressed 3rd person reflexive pronoun sei, soi is partly replaced by the ordinary personal pronoun, so that, in Modern “rench, soi is restricted to the role of reflexive marker for indefinite subjects. This leads to a partial identity of reflexive and intensifier, as both contain the pronoun, e. g. avec lui ୸ lui-même. ”erman shows some variation between the written standard and regional varieties. In the 1st and 2nd persons, reflexivity is expressed by personal pronouns in Old High ”erman, while regionally, sich can be found today. In the 3rd person, the dative of the reflexive pronoun is lost in Old High ”erman, but replaced by the accusative in Early Modern High ”erman, although there are regional exceptions. These developments disprove Vennemann’s claim that the expression of reflexivity remained unchanged for 5000 years in ”erman and Romance. Both North Sea ”ermanic and Insular Celtic lack PIE *s( )e- altogether, and both language branches initially used personal pronouns to fill the gap. The individual languages then developed different strategies to resolve the resulting cases of ambiguity. 4.2 The rise of complex intensifiers Table 8 summarizes the rise of complex intensifiers in Welsh, English and Irish. Complex intensifiers are present in the earliest Old British attestations, while in English and Irish they develop only during the middle periods. Based on this chronology, one could speculate that the feature spread from Old British or Welsh into English, although it remains difficult to propose a convincing time-frame and socio-historic context for this development. Complex intensifiers are attested in Old British from the early 9th century onwards. Their appearance could be dated around a century earlier, due to the fact that identical intensifiers are found in Old South-West-British. There is however no way to know if complex intensifiers are even older and if they emerged in Old English due to language contact between Britons and Anglo-Saxons during the 5th and 6th centuries. Tristram (1999: 16, 29) claims that Old English acquired most of its Celtic features during this period, when the indigenous population rapidly shifted from Brittonic to English. However, the feature could have also spread at a later time, as language contact between Welsh and English continued. Vezzosi (2005: 236) points out that the genitive pattern of complex intensifiers is concentrated in Middle English texts from the West-Midlands (See also Lange 2007: 61). Table 8. The rise of complex intensifiers in Brittonic, English and Irish (± indicates the co-occurrence of alternative strategies) MW 1150–1400 Early ModW 1400– + + + OE 650–900 OE–ME 1000–1500 Early ModE 1500– + OIr 750–900 MIr 900–1200 Early ModIr 1600– ± + + ± + Intensifiers and reflexives in SAE, Insular Celtic and English simple intensifier poss. pron. + intensifier pers. pron. + intensifier pers./poss. pr. + intens. Old Brit. 800–1150 197 198 Britta Irslinger During the same period, complex intensifiers emerge also in Irish. This has however barely been taken into consideration by the scholars who discussed their emergence through language contact. ’omplex intensifiers can of course also be found in the neighbouring languages and are thus not as exceptional as it has sometimes been claimed with regard to Insular ’eltic and English. In Vulgar Latin, intensifiers are composed of personal pronouns suffixed with -met, to which ipse self is added, resulting in egometipse (I) myself , tumetipse (you) yourself , illemetipse (he) himself , semetipse etc.47 The superlative *metipsimus developed into Old “rench méìsme, mesme, Modern “rench même. This intensifier is also combined with personal pronouns, cf. Old “rench moy mesme, lui medisme, Modern “rench moi-même myself , lui-même himself etc. (Rheinfelder 1976: 177). 4.3 Complex intensifiers become reflexive markers More remarkable than the rise of complex intensifiers is their use as reflexive markers. This change occurs in English, Welsh and Irish roughly contemporaneously. Although examples of complex intensifiers used as reflexives can be found already in the middle periods, the grammaticalisation of this use happens decidedly later than the emergence of the complex intensifiers themselves (Table 7). Lange (2007: 176) stated this for English, where the development can be dated precisely. In Welsh and Irish, more research is needed. An extensive corpus showing the distribution of the different constructions has not yet been compiled. Until this much needed evidence is collected and evaluated some of the above conclusions can only be taken as provisional. In Irish, the development takes place after the middle period and is probably accomplished before the early modern stage. As in English, complex intensifiers are used as reflexives for 3rd persons earlier than for 1st and 2nd persons. In Welsh, the situation is more complicated. ’omplex intensifiers are used as reflexives already in Middle Welsh, but they compete with reflexive ym-verbs. It seems that ei hun-reflexives started to replace ym-verbs only around 1550. Whether they were frequent enough to exert any influence on English at an earlier time remains open to debate. At present, it seems safer to assume mutual convergence. Moreover, the possibility of parallel developments cannot be excluded, and several scholars explain the emergence of the English reflexives on 47 See Hofmann & Szantyr 1965: 174; Väänänen 1981: 123, § 279; Pieroni 2010: 461. Intensifiers and reflexives in SAE, Insular Celtic and English 199 Table 9. The use of complex intensifiers as reflexive pronouns (± indicates the co-occurrence of alternative strategies) Language stage Old British MW Late ModW Late Corn. OE OE-ME Early ModE OIr MIr Early ModIr 800–1150 1150–1400 1550– 1600–1750 650–900 1000–1500 1500– 750–900 900–1200 1600– – ± ± + – ± + – ± + the basis of language internal developments.48 This of course applies to Welsh and Irish as well. Nevertheless, attention should be drawn to a piece of evidence which has not been considered before, namely the respective developments in Cornish and Breton. Both languages have the same intensifiers as Welsh from the beginning of their attestation, which they do not, however, use as reflexives. Only Late Cornish, by 1600, abandons its inherited reflexivisation strategy in favour of possessive pronouns or complex intensifiers like Welsh and English. Cornish thus adopts the strategy used by its neighbouring languages. This may be connected to the fact that by that time Cornish had already become an endangered language (which died before the end of the 18th century). Breton, on the contrary, maintains em as a reflexive marker until present. As this marker is no longer a preverb but a particle , it has become more similar to the “rench reflexive pronouns. As its intensifiers have not changed, Breton possesses the intensifier-reflexive differentiation just as much as its neighbours. The same is true for the West ”ermanic languages Dutch, Afrikaans and “risian, which have lost the reflexive *s( )e-. Dutch has reacquired a reflexive by borrowing. Afrikaans created a new complex reflexive. Only “risian uses simple personal pronouns as reflexives, but it has complex reflexives for emphatic use. The diachronic developments have thus led to the emergence of two distinct convergence areas, which both contain Insular Celtic and West ”ermanic languages. All languages on the Isles use complex intensifiers as reflexives, while the languages on the Continent possess the intensifier-reflexive differentiation. 48 See e. g. König & Siemund 2000; Bergeton & Pancheva 2012. 200 Britta Irslinger 4.4 Semitic influence on Old British intensifiers? It has been shown that the rise of complex intensifiers and their use as reflexives are two independent developments, the latter of which occurs only after the turn of the first millennium. It has also been shown that this development takes place roughly contemporaneously in Irish, Welsh and English, disproving the assumption that Insular ’eltic languages used this reflexivisation strategy (much) earlier than English. Substratal influence could be involved at most in the rise of complex intensifiers, since they are attested in Old ‘ritish before they appear in Irish or Welsh. Several scholars portray the Welsh complex intensifiers as exceptional in the context of European languages. However, the reflexive in ‘asque is a phrase consisting of possessive + head , and in ‘rabants Dutch it is possessive + own . It has also been shown that combinations of possessive + body are attested in Old and Middle High German, Old French and Middle English. This list is not exhaustive, and an analysis that covers more languages could likely reveal more constructions of this type, as they are cross-linguistically widespread and seem to rise independently or by borrowing. ’onsequently, ascribing the derivation of the Old ‘ritish complex intensifiers to Phoenician influence is not a claim to be accepted without adequate evidence. Acknowledgement: The author would like to thank the anonymous reviewer for the valuable comments that greatly contributed to improving the final version of the paper. Cited texts Ælfric, ”en. Ælfric, ”r. Alexius BM Catwn Chaucer, CT Samuel J. Crawford (1922). The Old English Version of the Heptateuch. Aelfric’s Treatise on the Old and New Testament and His Preface to Genesis. London: Milford. Julius Zupitza (1966). Aelfrics Grammatik und Glossar. Hildesheim: Weidmann. Chanson of Alexius from the St Albans Psalter. Transcription and English Translation cf. http://www.abdn.ac.uk/stalbanspsalter/english/translation/trans057. shtml. Whitley Stokes (1872). Beunans Meriasek. The life of Saint Meriasek, bishop and confessor. A Cornish drama. London: Trübner. Henry Lewis (1923a). Catwn a’i Ddehongliad . Peniarth Ms. 3, ii; p. 34. In: Bulletin of the Board of Celtic Studies 2, 8–36. “red N. Robinson (1957). The Works of Geoffrey Chaucer. Canterbury Tales, The Man of Law’s Introduction, Prologue and Tale. Oxford: Oxford University Press. Intensifiers and reflexives in SAE, Insular Celtic and English CW Cyngorau ”ower, CA ”wenolé ”wyrthyeu Heinrich IAI ILCV Isembart Iwein JKS KL MC Ml. MP Orosius Otfrid PD Peredur PH 201 Paula Neuss (1983). The Creacion of the World. A critical edition and translation. New York & London: ”arland. Henry Lewis (1923b). Testunau. Cyngorau Catwn . Llanstephan collection, National Library of Wales 27, 164b. In: Bulletin of the Board of Celtic Studies 2, 8–36. Russel A. Peck (2000–2005). John Gower: Confessio Amantis. 3 vols. Kalamazoo, MI: Medieval Institute Publications. Paul Widmer & Anders R. Jørgensen (2011). An Buhez Sant Gwenôlé. Das Leben des heiligen Gwenole. Text, Übersetzung und Anmerkungen. Wien: Edition Praesens. ”wenan Jones (1939–1941). ”wyrthyeu y Wynvydedic Veir . In: Bulletin of the Board of Celtic Studies 9.4, 144–148, 334–341 & 10, 21–33. Hartmann von Aue (2010). Der arme Heinrich. Ed. by Hermann Paul. 18., unveränderte Auflage neu bearbeitet von Kurt ”ärtner. Berlin & New York: de ”ruyter. Aotrou Kerne (1893). Istor hor Mamm zantel an Iliz. Landerneau: Desmoulins. Ernestvs Diehl (1961–1967). Inscriptiones Latinae Christianae Veteres. 2nd ed. 3 vols. Berlin: Weidmann. Karl Voretzsch (1966). Isembart and ”ormont . In: Altfranzösisches Lesebuch. 3rd ed. Tübingen: Niemeyer. Hartmann von Aue (2001). Iwein. Text der siebenten Ausgabe von ”. “. Benecke, K. Lachmann und L. Wolff. Übersetzung und Nachwort von Thomas Cramer. 4th ed. Berlin & New York: de ”ruyter. Amable-Emmanuel (c. 1862). Jezuz-Krist, Skouer ar Gristinien. Troude and ”ab Milin. Translation of the Imitatio Christi. Kynniuer llith a ban (1551). Early English Books online. Whitley Stokes (1873). The Old Welsh glosses in Martianus Capella: De Nuptiis Philologiae et Mercurii . Corpus Christi College, Cambridge MS 153. In: Beiträge zur vergleichenden Sprachforschung auf dem Gebiete der arischen, celtischen und slawischen Sprachen 7.4, 385–416. The Milan glosses on the psalms (1975b). In: Thesaurus Palaeohibernicus. A collection of Old Irish glosses and scholia prose and verse. Ed. by Whitley Stokes. Repr. Vol. 1. Dublin: Dublin Institute for Advanced Studies, 7ff. Pierre-Yves Lambert (2003a). The Old Welsh glosses on weights and measures . De mensuris et ponderibus. Liber Commonei, Oxoniensis Prior. Bodleian MS. Auct. “. 4. 32. In: Yr Hen Iaith. Studies in early Welsh. Ed. by Paul Russell. Aberystwyth: Celtic Studies Publications, 103–134. Henry Sweet (1883). King Alfred’s Orosius. Vol. 1: Old-English and Latin Original. London: Trübner. Oskar Erdmann (1973). Otfrids Evangelienbuch. 6th ed. Tübingen: Niemeyer. Edwin Norris (1968a). The Ancient Cornish Drama. Vol. 1: Passio domini nostri Jhesu Christi. Reprint of the edition Oxford 1859. New York & London: Benjamin Blom. ”lenys Witchard ”oetinck (1976). Historia Peredur vab Efrawc. Caerdydd: ”wasg Prifysgol Cymru. Robert Atkinson (1887). The Passions and Homilies from Leabhar Breac. Dublin: Academy. 202 PKM RD Rol. Schwein. Sg. Shakespeare Taith C Togail Troí Wb. William Rowe WM YBH YL YS” Yvain Britta Irslinger Ifor Williams (1930). Pedeir keinc y Mabinogi. Allan o Lyfr Gwyn Rhydderch. Caerdydd: ”wasg Prifysgol Cymru. Edwin Norris (1968b). The Ancient Cornish Drama. Vol. 2: Ressurectio Domini. Reprint of the edition Oxford 1859. New York & London: Benjamin Blom. Joseph Bédier (1964). La Chanson de Roland. Publiée d’après le manuscrit d’Oxford et traduite par J. B. Édition définitive. Paris: Piazza. Johann ”. ”. Büsching (1820). Lieben, Lust und Leben der Deutschen des 16. Jahrhunderts, in den Begebenheiten des Schlesischen Ritters Hans von Schweinichen, von ihm selbst aufgesetzt. Breslau: Josef Max. ”losses on Priscian (St. ”all) (1975a). In: Thesaurus Palaeohibernicus. A Collection of Old Irish Glosses and Scholia Prose and Verse. Ed. by Whitley Stokes. Repr. Vol. 2. Dublin: Dublin Institute for Advanced Studies, 49ff. Shake-speares sonnets (1609). Neuer before imprinted. At London: By ”. Eld for T[homas] T[horpe] and are to be solde by Iohn Wright, dwelling at Christ Church gate. (Early English Books online). Siwrneu, neu daith Cristiana a’i phlant o ddinas destryw. Togail Troí (1964–1983). In: The Book of Leinster. Ed. by Richard I. Best & Micheal O’Brien. Vol. 4. Dublin: Dublin Institute for Advanced Studies, 1063– 1117. The Würzburg glosses on the Pauline epistles (1975c). In: Thesaurus Palaeohibernicus. A Collection of Old Irish Glosses and Scholia Prose and Verse. Ed. by Whitley Stokes. Repr. Vol. 1. Dublin: Dublin Institute for Advanced Studies, 7ff. Translation of Matthew 4 by William Rowe (Cornish Willow Kereve, 1660–c. 1700). Transcription and English translation. http://www.bibelkernewek.com/ wrm4.htm (visited on 03/17/2013). J. ”wenogvryn Evans (1907). The White Book Mabinogion. Welsh tales and romances reproduced from the Peniarth manuscripts. Pwllheli. Morgan Watkin (1958). Ystorya Bown o Hamtwn. Cyfieithiad canol y drydedd ganrif ar ddeg o La geste de Boun de Hamtone. Caerdydd: ”wasg Prifysgol Cymru. John Morris Jones & John Rhŷs (1894). Ystoria Lucidar = Hystoria Lucidar. The Elucidarium and Other Tracts in Welsh from Llyvyr Agkyr Llandewivrevi. Oxford: Clarendon, 3–76. Thomas Jones (1992). Ystoryaeu Seint Greal. Vol. 1: Y keis. Caerdydd: ”wasg Prifysgol Cymru. Chrétien de Troyes. Le chevalier au lion ou le roman d’Yvain. Manuscrit Paris, B.N. fr. 794 (ms. H). Wilhelm L. Holland (1880). Li romans dou Chevalier au Lyon von Crestien von Troies. 2nd ed. Hannover: Rümpler 1880. See also http://www.uottawa.ca/academic/arts/lfa/activites/textes/chevalierau-lion/H/Hpresentation.html. Intensifiers and reflexives in SAE, Insular Celtic and English 203 Bibliography Beekes, Robert S. P. (1983). On laryngeals and pronouns . In: Zeitschrift für vergleichende Sprachforschung 96, 200 232. (2011). Comparative Indo-European Linguistics. An Introduction. 2nd ed. Amsterdam & Philadelphia: Benjamins. Behagel, Otto (1923). Deutsche Syntax. Eine geschichtliche Darstellung. Vol. 1: Die Wortklassen und Wortformen. A: Nomen. Pronomen. Heidelberg: Winter. Bergeton, Uffe & Roumyana Pancheva (2012). A new perspective on the historical development of English intensifiers and reflexives . In: Grammatical Change. Origins, Nature, Outcomes. Ed. by Dianne Jonas, John Whitman & Andrew ”arrett. Oxford: Oxford University Press, 123 138. Besch, Werner (1967). Sprachlandschaften und Sprachausgleich im 15. Jahrhundert. Studien zur Erforschung der spätmittelhochdeutschen Schreibdialekte und zur Entstehung der neuhochdeutschen Schriftsprache. München: “rancke. Borsley, Robert D., Maggie Tallerman & David Willis (2007). The Syntax of Welsh. Cambridge: Cambridge University Press. Braune, Wilhelm & Ingo Reiffenstein (2004). Althochdeutsche Grammatik. Vol. 1: Laut- und Formenlehre. 15th ed. Tübingen: Niemeyer. Breatnach, Liam (1994). An Mheán-”haeilge . In: Stair na Gaeilge in ómós do Pádraig Ó Fiannachta. Ed. by Kim McCone et al. Maigh Nuad: Roinn na Sean-”haeilge, 221 333. Brugmann, Karl (1911). Grundriss der vergleichenden Grammatik der indogermanischen Sprachen. Vol. 2: Lehre von den Wortformen und ihrem Gebrauch. 2. Zahlwörter. Die drei Nominalgenera. Kasus- und Numerusbildung der Nomina. 2nd ed. Strassburg: Trübner. Cennamo, Michaela (1991). Se, sibi, suus nelle Inscriptiones Latinae Christianae Veteres ed i successivi sviluppi romanzi . In: Medioevo Romanzo 16, 3 20. Clackson, James (2007). Indo-European Linguistics. An Introduction. Cambridge & New York: Cambridge University Press. Cotticelli Kurras, Paola & Alfredo Rizza (2013). Reconstructing Proto-Indo-European categories. The reflexive and the middle in Hittite and in the Proto-language . In: Journal of Historical Linguistics 3.1, 7 27. De Lage, ”uy R. (1990). Introduction à l’ancien français. Nouv. éd. par ”eneviève Hasenohr. Paris: SEDES. Dottin, ”eorges (1987). Manuel d’irlandais moyen. Grammaire, textes et glossaire. Repr. 1913. ”enève: Slatkine Reprints. Dunkel, ”eorge E. (2014). Lexikon der indogermanischen Partikeln und Pronominalstämme. Vol. 2: Lexikon. Heidelberg: Winter. Ebert, Robert Peter et al. (1993). Frühneuhochdeutsche Grammatik. Tübingen: Niemeyer. Evans, D. Simon (1964). A Grammar of Middle Welsh. Dublin: Dublin Institute for Advanced Studies. “alileyev, Aleksander (2000). Etymological Glossary of Old Welsh. Tübingen: Niemeyer. “alileyev, Alexandre I. (2008). Le Vieux-Gallois. Ed. by Hildegard L. C. Tristram. Potsdam: Universitätsverlag. “avereau, “rancis (1997). Grammaire du Breton contemporain / Yezhadur ar brezhoneg a-vremañ. Morlaix: Skol Vreizh. “leuriot, Léon (1964a). Dictionnaire des glosses en vieux breton. Paris: Klincksieck. (1964b). Le vieux breton. Éléments d’une grammaire. Paris: Klincksieck. 204 Britta Irslinger “lobert, Pierre (1975). Les verbes déponents latins des origines à Charlemagne. Paris: Belles Lettres. “rings, Theodor (1957). Grundlegung einer Geschichte der deutschen Sprache. Halle: Niemeyer. ”rimm, Jacob (1866). Kleinere Schiften. Vol. 3. Berlin: Dümmler. Hackstein, Olav (2003). Reflexivpronomina, Präverbien und Lokalpartikel in indogermanischen Sprachen . In: Tocharian and Indo-European Studies 10, 69 95. Harbert, Wayne (2007). The Germanic Languages. Cambridge: Cambridge University Press. Haspelmath, Martin (2001). The European linguistic area. Standard Average European . In: Language Typology and Language Universals. Ed. by Martin Haspelmath et al. Vol. 2. Berlin & New York: de ”ruyter, 1492 1510. (2003). The geometry of grammatical meaning: semantic maps and cross-linguistic comparison . In: The New Psychology of Language. Ed. by Michael Tomasello. Vol. 2. Mahwah, NJ: Erlbaum, 211 242. Heine, Bernd (2005). On reflexive forms in creoles . In: Lingua 115, 201 257. Heine, Bernd & Tania Kuteva (2002). World Lexicon of Grammaticalization. Cambridge: Cambridge University Press. Hofmann, Johann B. & Anton Szantyr (1965). Lateinische Grammatik. Vol. 2: Lateinische Syntax und Stilistik. München: Beck. Keenan, Edward L. (1996). Creating Anaphors. A Historical Study of English Reflexive Pronouns. Ms. Kemmer, Suzanne (1993). The Middle Voice. Amsterdam: Benjamins. King, ”areth (2003). Modern Welsh. A Comprehensive Grammar. 2nd ed. Comprehensive grammars. London: Routledge. Klingenschmitt, ”ert (1994). Die Verwandtschaftsverhältnisse der indogermanischen Sprachen . In: In memoriam Holger Pedersen. Kolloquium der Indogermanischen ”esellschaft vom 26. bis 28. März 1993 in Kopenhagen. Ed. by Jens E. Rasmussen. Wiesbaden: Reichert, 235 251. Kluge, “riedrich & Elmar Seebold (2011). Etymologisches Wörterbuch der deutschen Sprache. 25th ed. Berlin & New York: de ”ruyter. König, Ekkehard (2001). Intensifiers and reflexive pronouns . In: Language Typology and Language Universals. Ed. by Martin Haspelmath et al. Berlin & New York: de ”ruyter, 747 760. König, Ekkehard & Volker ”ast (2009). Understanding English-German Contrasts. Berlin: Erich Schmidt. König, Ekkehard & Peter Siemund (2000). The development of complex reflexives and intensifiers in English . In: Diachronica 17.1, 39 84. Lambert, Pierre-Yves (2003b). The Old Welsh glosses on weights and measures . In: Yr Hen Iaith. Studies in early Welsh. Ed. by Paul Russell. Aberystwyth: Celtic Studies Publications, 103 134. Lange, Claudia (2005). Reflexivity and intensification in Irish English and other new Englishes . In: The Celtic Englishes IV. The Interface between English and the Celtic Languages. Proceedings of the fourth International Colloquium on the Celtic Englishes held at the University of Potsdam in ”olm (”ermany) from 22 26 September 2004. Ed. by Hildegard L. C. Tristram. Potsdam: Universitätsverlag, 259 282. (2007). Reflexivity and Intensification in English. A Study of Texts and Contexts. “rankfurt: Lang. Le Roux, Pierre (1957). Le verbe breton (morphologie, syntaxe). 2nd ed. Rennes and Paris: Plihon and Champion. Intensifiers and reflexives in SAE, Insular Celtic and English 205 Lewis, Henry & Stefan Zimmer (1990). Handbuch des Mittelkornischen. Deutsche Bearbeitung von Stefan Zimmer, mit einem Anhang von Andrew Hawke. Innsbruck: Institut für Sprachwissenschaft. Lutz, Angelika (2002). The reflexive in Middle English. Loose ends in the pronominal system . In: Middle English from Tongue to Text. Selected papers from the Third International Conference on Middle English. Language and text, held at Dublin, Ireland, 1 4 July 1999. Ed. by Peter J. Lucas & Angela Lucas. Bern: Lang, 45 68. McCone, Kim (2006). The Origins and Development of the Insular Celtic Verbal Complex. Maynooth: Department of Old Irish. Meiser, ”erhard (2009). Zur Typologie des urindogermanischen Mediums . In: Protolanguage and Prehistory. Akten der XII. “achtagung der Indogermanischen ”esellschaft, Krakau, 11. bis 15. Oktober 2004. Ed. by Rosemarie Lühr & Sabine Ziegler. Wiesbaden: Reichert, 318 334. Miller, D. ”ary (1993). Complex Verb Formation. Amsterdam: Benjamins. Morris Jones, John (1913). A Welsh Grammar. Historical and Comparative. Oxford: Clarendon. Mustanoja, Tauno “. (1960). A Middle English Syntax. Vol. 1: Parts of Speech. Helsinki: Soc. néophilologique. Nolan, Brian (2012). The Structure of Modern Irish. A Functional Account. Sheffield: Equinox. Ó Curnáin, Brian (2007). The Irish of Iorras Aithneach, County Galway. 4 vols. Dublin: Dublin Institute for Advanced Studies. Ó Dónaill, Niall, ed. (1977). Foclóir Gaeilge-Béarla. Baile Átha Cliath: Oifig an tSoláthair. Ó Siadhail, Mícheál (1989). Modern Irish. Grammatical Structure and Dialectal Variation. Cambridge: Cambridge University Press. Parina, Elena (2007). Reflexivpronomina im Mittelkymrischen . In: Kelten-Einfälle an der Donau. Akten des 4. Symposium deutschsprachiger Keltologinnen und Keltologen (Linz/Donau, Juli 2005). Ed. by Helmut Birkhan. Wien: Akademie der Wissenschaften, 389 396. Paul, Hermann (1919). Deutsche Grammatik. Vol. 3.4: Syntax (1. Hälfte). Halle: Niemeyer. (2007). Mittelhochdeutsche Grammatik. Ed. by Thomas Klein. 25th ed. Tübingen: Niemeyer. Peitsara, Kirsti (1997). The development of reflexive strategies in English . In: Grammaticalization at Work. Studies of Long-Term Developments in English. Ed. by Matti Rissanen, Merja Kytö & Kirsti Heikkonen. Berlin & New York: Mouton de ”ruyter, 277 370. Petit, Daniel (2001). Linguistique comparative et variation typologique: le cas du réfléchi dans les langues indo-européennes . In: Indogermanische Forschungen 106, 16 29. Picoche, Jacqueline & Christiane Marchello-Nizia (1998). Histoire de la language française. 5. éd. revue et corrigée. Paris: Éditions Nathan. Pieroni, Silvia (2010). Deixis and anaphora . In: New Perspectives on Historical Latin Syntax. Vol. 3: Constituent Syntax: Quantification, Numerals, Possession, Anaphora. Ed. by Philip Baldi & Pierluigi Cuzzolin. Berlin & New York: de ”ruyter, 389 501. Pilch, Herbert (1996). Word-formation in Welsh and Breton. A comparative study . In: Zeitschrift für celtische Philologie 48, 34 88. Pinault, ”eorges-Jean (2013). Body and soul: the reflexive in Tocharian . In: Indogermanische Forschungen 118, 339 359. Poppe, Erich (2009). Standard Average European and the Celticity of English intensifiers and reflexives: some considerations and implications . In: English Language and Linguistics 13.2, 251 266. Poppe, Erich & Regine Reck (2008). Rewriting Bewis in Wales and Ireland . In: Sir Bevis of Hampton in Literary Tradition. Ed. by Jennifer “ellows & Ivana Djordjević. Woodbridge: Brewer, 37 50. 206 ‘ritta Irslinger Postma, ”ertjan (2012). Language contact and linguistic complexity the rise of the reflexive pronoun zich in a fifteenth-century Netherlands border dialect . In: Grammatical Change. Origins, Nature, Outcomes. Ed. by ‚ianne Jonas, John Whitman & Andrew ”arrett. Oxford: Oxford University Press, 139 159. Puddu, Nicoletta (2005). Riflessivi e intensificatori. Greco, latino e le altre lingue indoeuropee. Pisa: Ed. ETS. Rheinfelder, Hans (1976). Altfranzösische Grammatik. Vol. 2: Formenlehre. 2nd ed. München: Hueber. Rijk, Rudolf P. ”. de (2008). Standard Basque. A Progressive Grammar. Prepared by Armand ‚e ’oene. 2 vols. ’ambridge, Mass.: MIT Press. Schrijver, Peter (1997). Studies in the History of Celtic Pronouns and Particles. Maynooth: ‚epartment of Old Irish. (2011). Old ‘ritish . In: Brythonic Celtic – Britannisches Keltisch. Ed. by Elmar Ternes. ‘remen: Hempen, 1 84. Schumacher, Stefan (2004). Die keltischen Primärverben. Ein vergleichendes, etymologisches und morphologisches Lexikon. Unter Mitarbeit von ‘ritta Schulze-Thulin und ’aroline aan de Wiel. Innsbruck: Institut für Sprachen und Literaturen. (forthc.). y(d) und yt: Zur Syntax zweier Partikeln des Mittelkymrischen . In: Keltische Forschungen 6. Stéfanini, Jean (1962). La voix pronominale en ancien et en moyen français. ”ap: Ophrys. Thomas, Peter W. (1996). Gramadeg y Gymraeg. ’aerdydd: ”wasg Prifysgol ’ymru. Tichy, Eva (2004). Indogermanistisches Grundwissen für Studierende sprachwissenschaftlicher Disziplinen. 2., überarbeitete Auflage. ‘remen: Hempen. Trask, R. L. (2003). Morphology . In: A Grammar of Basque. Ed. by José Ignacio Hualde & Jon Ortiz de Urbina. 2 vols. ‘erlin & New York: Mouton de ”ruyter, 113 186. Tristram, Hildegard L. ’. (1999). How Celtic is Standard English? St. Peterburg: Nauka. Väänänen, Veikko (1981). Introduction au latin vulgaire. 3. éd. rev. et augm. Paris: Klincksieck. Van ”elderen, Elly (2000). A History of English Reflexive Pronouns. Person, Self, and Interpretability. Amsterdam: ‘enjamins. Vendryes, Joseph (1927). Les verbes composés avec ym- dans les Mabinogion . In: Mélanges bretons et celtiques offerts à M. J. Loth. Rennes & Paris: Plihon & ’hampion, 49 62. Vennemann, Theo (2013). ’oncerning myself . In: Lexical and Structural Etymology. Beyond Word Histories. Ed. by Robert Mailhammer. ‘oston & ‘erlin: de ”ruyter, 121 146. Vezzosi, Letizia (2005). The development of himself in Middle English. A ’eltic Hypothesis . In: Rethinking Middle English. Linguistic and Literary Approaches. Ed. by Nikolaus Ritt & Herbert Schendl. “rankfurt am Main: Lang, 228 243. Vries, Jan de (1962). Altnordisches etymologisches Wörterbuch. 2nd ed. Leiden: ‘rill. Williams, Ifor (1931). ”losau Rhydychen: Mesurau a phwysau . In: Bulletin of the Board of Celtic Studies 5, 226 248. Wmffre, Iwan (1998). Late Cornish. München: Lincom. Zimmer, Stefan (2000). Studies in Welsh Word-Formation. ‚ublin: School of ’eltic Studies, ‚ublin Institute for Advances Studies.