Simplification in Derivational Morphology in the Lindisfarne Gloss

1 Luisa García García Simplification in Derivational Morphology in the Lindisfarne Gloss Copyright: This chapter is published in The Old English Gloss to the Lindisfarne Gospels: Language, Author and Context, edited by Fernández Cuesta, Julia and Pons-Sanz, Sara M., 2016, pp. pp. 189-212, Berlin, Boston: De Gruyter, reproduced with permission of the publisher. Abstract: The language of the English gloss to the Lindisfarne Gospels bears witness to early morphological simplification in the late Northumbrian dialect compared to other Old English varieties. Cases of merger in the nominal inflection have already been widely noted. Morphological syncretism in the area of derivation has attracted less attention, although here too variation with respect to other Old English dialects is to be expected. The aim of this paper is to establish how the language of the Lindisfarne gloss differs from that of other Old English texts concerning the degree of syncretism of the causative jan-formation. To that end, all deverbal jan-pairs (base and derivative) in the gloss have been identified, and the syntactic valency and meaning of each member of the pair have been assessed by studying each attestation individually. The main conclusion reached is that the language of the Lindisfarne gloss does not show more innovative traits than that of other Old English texts in the use of the causative formation. This points to derivational morphology behaving differently from inflexional morphology with respect to morphological loss in Old Northumbrian. 1 Introduction: the jan-formation in early Germanic languages In this section, the Germanic causative formation will be briefly described, differences in the lexicalization stage of causatives among West Germanic languages will be pointed out and the aim of this paper will be more narrowly defined against that background. Section 2 focuses on Old English causatives and their processes of semantic and morpho-syntactic change. Section 3 explains the methodology followed in this study, its sources and the process of data collection. Section 4 presents the data and the results obtained. Section 5 presents the conclusions of the study and the questions which deserve further research. An Appendix contains a list of all the jan-causatives and their derivational bases attested in the Lindisfarne gloss, with an indication of their meaning and syntactic use in the gloss as well as in ‘general’ Old English (for this term see note to section 2 below). A label indicates the relationship between janand strong verb in Lindisfarne, general Old English, and Germanic. The selection of items in the list and the information contained there have resulted from research conducted for this paper and from previous work (see García García 2012a). All jan-pairs discussed below may be found in the Appendix. In the Germanic proto-language there existed a productive word-formation mechanism for deriving causative verbs from non-causative verbal bases by means of the ija-suffix. It is the most common deverbal word-formation pattern in this language, where deverbal jan-pairs constitute a significant portion of the verbal lexicon, with roughly a third of all Germanic strong verbs (about 640 listed by Seebold 1970) having a jan-derivate attested in one or another Germanic language. This mechanism, which goes back to Indo-European, consists of the addition of the aforementioned ija-suffix to a primary verbal root in the Indo-European ograde (Germanic a-grade). One example is: 2 (1) PGmc *seta- ‘to sit, be seated’ strong base *sat-ija- ‘to place, set, sit (sth.)’ weak derivative. Although this pattern had ceased to be productive by the time of the first Germanic texts, jancausatives constitute a substantial part of the verbal lexicon of early Germanic languages and are still recognizable in Present-Day English; cp. sit  set; fall  fell; rise  rear; drink  drench; lie  lay. As has been pointed out in previous research (García García 2012b, 2013), the relationship between the base verb and its jan-derivative is formally rather opaque in all Germanic languages except Gothic, and is further obscured by semantic and syntactic changes associated with a process of lexicalization understood literally as a shift from grammar to lexicon. 1 As an example of semantic change, in this case metaphorization, undergone by both members of the jan-pair consider: (2) a. non-causative base OE weallan ‘to bubble forth, flow; well (with); exist in large numbers’ b. causative derivate OE wyllan ‘to boil (something); torment, agitate (someone)’. The causative jan-pair PGmc *melta- ‘to melt’ (non-causative/intrans.) ~ *maltija- ‘to melt’ (causative/trans.) provides a good example of syntactic change. Whereas in Germanic the causative and non-causative senses of the verb melt are distinguished morphologically, in Present-Day English both senses are expressed by the same form, in what is known as a “labile causative opposition” (Haspelmath 1993: 91, and n. 7): (3) The ice melts (non-causative) The sun melts the ice (causative). Note that causativization is a valency-increasing word-formation mechanism, that the causative derivatives of intransitive verbs differ from their transitive counterparts only in syntactic valency, as in for example fall  fell. 2 Old English represents a transition between Proto-Germanic and Present-Day English, as in the following example: (4) a. causative OE myltan (wk I) ‘to melt’ (trans. and intr.) < PGmc *maltija- ‘melt’ (trans.) b. non-causative OE meltan (st. 3) ‘melt’ (mostly intrans.) < PGmc *melta- ‘melt’ (intrans.). 1 See for instance Brinton and Traugott (2005: 54), quoting Ramat (1992: 557): “today’s grammar may become tomorrow’s lexicon”. Incidentally, Brinton and Traugott refer repeatedly to Germanic causatives as instances of lexicalization (see pp. 54, 87, 105, 153 among others). 2 For a detailed description of valency-changing morphological mechanisms cross-linguistically, see Haspelmath and Müller-Bardey (2000: 1130‒1145). 3 The weak verb, originally causative, is labile, that is to say, it can be used in both a causative and an intransitive sense in Old English. However, the intransitive use was originally reserved to the non-causative strong base. It is a case of ‘syntactic merger’ in that the valency frames of the causative and its base are no longer kept apart, and the morphological alternation myltan (wk) ~ meltan (st.) is functionally empty (see García García 2012a: 137‒139). Predictably, once syntactic merger has begun, one (or both) of the members of the causative opposition will disappear or change its (their) meaning. The loss of the distinctions formerly expressed by a given word-formation pattern, in this case the causative jan-formation, constitutes an instance of morphological simplification at the derivational level. As a result of the lexicalization process of jan-causatives, the causative sense might ultimately become barely recognizable, as in singe, the jan-causative derived from sing, or disappear entirely, as in OE onegan ‘to fear’, jan-derivative to og ‘fear’; cp. Go. jan-verb ogjan ‘to scare (someone)’. 3 As explained in García García (2013: 253‒256), jan-causatives are affected by semantic and syntactic changes to a greater extent in Old English than in other Germanic and even West Germanic languages. Old English causatives exhibit a higher degree of lexicalization. Thus, for instance the Go. ogjan ‘to scare (someone)’ above preserves its causative sense, but OE onegan ‘to fear’ does not. By comparing their nominal and verbal paradigms it is apparent that Old English is inflexionally less rich than other West Germanic languages, with the exception of Old Frisian. A well-known example is the plural present inflexion, with three different personal endings in Old High German against one in Old English (Old Saxon and Old Frisian coincide with Old English in this). The nominal inflection provides further examples, as detailed in García García (2000). In addition, note that Old Saxon seems to occupy an intermediate position between Old High German and Old English, with, for instance, overt instrumental case in a-, iand u-stems, like Old High German. The causative formation supplies an example of comparatively greater syncretism of Old English in the area of derivational morphology. Compare the reflexes of the Germanic causative pair *-leiba- ‘to remain’ ~ *-laibija- ‘to leave’ in Old High German, Old Frisian, Old Saxon, and Old English: (5) OHG biliban ‘to stay’ OHG leiben ‘to leave, leave unfinished’ OFris. biliva ‘to stay’ OFris. leva ‘to leave, let’ OS bilivan ‘to remain’ OS farlevian ‘to leave (over)’ OE belifan ‘to be left over, remain’ OE læfan, belæfan ‘to leave, remain’. The causative opposition expressed by this jan-pair has remained intact in all West Germanic languages except Old English. Both OE belifan and belæfan can express the (non-causative) meaning ‘to remain’, with OE belæfan adopting the valency frame of the non-causative verb OE belifan. This is, as mentioned above, a process of syntactic merger that leads to morphological indistinctiveness, as the same meaning (in this case ‘remain’) can be expressed by both the strong and the jan-verb. The example illustrates a process that is often alluded to, though there has been no systematic study of it to date. The morphological behaviour of Old Northumbrian with respect to other Old English varieties seems to replicate that of Old English – even at its most inflexionally complex, early West Saxon – with respect to other early Germanic languages. The early inflexional syncretism undergone by Northumbrian has 3 Not all jan-verbs are causative. For a description of the functions of the Germanic -ija-suffix, see Ringe (2006: 252–254). 4 been widely acknowledged (see Ross 1937: 119‒124; and standard Old English grammars, especially Brunner 1965 passim). 4 To mention just two examples, consider the disintegration of the gender system (Brunner 1965: 195 n.; Fernández Cuesta et al. 2008: 139) and analogical levelling towards the most common noun inflectional pattern, masculine a-stems (see Fernández Cuesta and Rodríguez Ledesma 2001a: 480–481). 5 The aim of this paper is to determine whether the equation holds for derivational morphology too, as illustrated by the causative formation in the Old English gloss to the Lindisfarne Gospels (hereafter: Lindisfarne). The meaning and valency of the jan-causatives and their bases attested in Lindisfarne will be established and contrasted with those found in other Old English varieties. The data that provide the base of comparison are drawn from a previous study on Old English causatives and will be summarized in the following section. 2 Old English causatives and their lexicalization process As pointed out in García García (2012a), the Old English lexicon contains 106 deverbal janverbs with potential causative meaning with respect to their strong bases. 6 In that article the meaning and syntactic use or valency of both the non-causative bases and the jan-derivatives (a total of 212 verbs) were determined in order to ascertain the current relationship between the base and its potentially causative derivative in Old English. The Dictionary of Old English (hereafter: DOE), complete up to letter G, and the nineteenth-century but still indispensable Old English dictionaries by Bosworth and Toller (1882–1898 and 1908–1921) and Clark Hall (1960; first edition in 1894) were consulted for every item. In cases where the meaning or valency of a formation is not sufficiently clear or different dictionaries supply conflicting versions, primary sources were collated via the Dictionary of Old English Web Corpus (hereafter: DOEC). Once the meaning and use of Old English jan-verbs with respect to their verbal bases were established, they were compared with those of their cognates in other Germanic languages, with the aim of determining the original derivational meaning of each pair in Proto-Germanic as the starting point from which that extant in Old English developed. The reconstruction of Proto-Germanic jan-pairs derives partly from previous work (García García 2005). As a result of the analysis, a basic list was compiled of Old English deverbal jan-verbs that were most probably causative in origin (including one which was intensive/iterative). The relationship between each jan-verb and its base, so far as it is attested, was analysed and labelled, both in Old English and in Proto-Germanic, when the jan-formations were coined. By comparing the function of jan-verbs in the two language stages it is possible to assess the changes that these verbs have undergone in Old English and identify some tendencies in their evolution, which were summarized in the previous section of the present paper. Table 1 below demonstrates in figures the development of all 106 deverbal jan-verbs in Old English. Fifty-seven of them have a causative sense with respect to their bases in Old English, as against 71 in Proto-Germanic. These have evolved into Old English as follows: 56 Germanic causative pairs remain causative in Old English, three are doubtful, four develop 4 Toon (1992), in his survey of Old English dialects for the first volume of The Cambridge History of the English Language, does not deal with it. 5 Extension of the endings of the n-declension to strong nouns does also occur in Old Northumbrian to a greater extent than in other varieties (Ross 1937: 101; Brunner 1965: 196). 6 In this paper, and in the coming sections in particular, the term ‘Old English’ refers to ‘undifferentiated’ or ‘general’ Old English, including Old Northumbrian and the language of the Lindisfarne gloss. Often, as in the choice of spelling for infinitive forms, (general) Old English is skewed towards the late West Saxon variety. However, in the meaning and syntactic description of verbs, all areas and periods of Old English are represented. 5 idiosyncratic semantic relationships, five show no difference between the members of the pair, and three Proto-Germanic causative pairs are clearly not causative in Old English. Table 1. Function of Old English deverbal jan-verbs with respect to their base, in figures 7 Causative 57 (56 causative in Germanic) Causative? 12 (3 causative in Germanic) Idiosyncratic 4 (4 causative in Germanic) No difference 20 (5 causative in Germanic) Non-causative 9 (3 causative in Germanic) Not identifiable 3 Intensive? 1 TOTAL 106 (71 causative in Germanic) Notice that of the 57 secure causative pairs attested in Old English (first line), nine show semantic deviation in one or both of their members; they are labelled ‘C+IDI’ (causative with idiosyncratic semantic development) and ‘C+SPE’ (causative with specialized meaning) in the mentioned article. Thirteen have changed their valency frame, and are labelled ‘ColC’ (collapsing causative opposition). This labelling convention is maintained in the final list in the appendix. The results of the study just described are global, that is, do not reflect textual, temporal or dialectal variation. This means that the deverbal jan-verbs that appear in Lindisfarne are obviously included, but they are not considered separately from other Old English texts. The next step is thus to contrast those global results against the textual data in Lindisfarne, as a representative of Old Northumbrian. The methodology of the study will be described next. 3 Methodology, sources and data collection process To carry out the present study, the occurrences of all 106 Old English deverbal jan-verbs and their strong bases attested in Lindisfarne were checked to begin with. The exact meaning and syntactic use of the relevant verbs were determined and the relationship between the jan-verb and its base was classified according to the types listed in Table 1 above. The results obtained for the jan-verbs attested in Lindisfarne were then contrasted with those obtained for the whole Old English period. For the purposes of the study (degree of morphological syncretism in Lindisfarne causatives), only those 57 verb-pairs for which a causative relation holds in Old English (see Table 1) were found relevant. They will be included in the final list attached. All the clauses in which they are attested in Lindisfarne, together with their Latin original, were analysed. The relationship between the members of the jan-pairs is labelled as ‘Causative’, ‘Causative+SPE’, ‘Causative+IDI’ and ‘ColCausative’, depending on whether they are straightforward causatives or display any semantic and/or syntactic (valency) changes (see above). In addition, one of the 12 instances of ‘Causative?’ (that is, a causative relation is 7 In brackets, the number of verbs in each group that has causative function in Germanic (adapted from García García 2012a: 135). 6 suspected but impossible to confirm) registered in Table 1 was clarified thanks to the present study. Concerning the sources of the primary data for the study, the localization of janderivatives and their basic strong verbs was carried out by means of comprehensive searches in the DOEC; the task was facilitated by Cook’s (1894) glossary and the DOE. Skeat’s edition of the glosses to the Lindisfarne Gospels was used primarily (Skeat 1871‒1887), mainly via the DOEC. However, the facsimile of the original manuscript was routinely consulted in the digitized version of the British Library webpage in all less than straightforward cases. For the purposes of the study both the jan-verb and its strong base are equally relevant. That is, if the jan-member of an Old English causative pair is not attested in Lindisfarne but its base is, the semantic and syntactic behaviour of the latter has to be analysed in order to determine whether it is additionally used in a causative sense (as a labile verb), in lieu of its jan-counterpart. Simple formations were chosen as primary data to avoid semantic interference by prefixes. If these were not attested for a given verb, all its prefixed formations were considered, starting with ge-formations, which are least likely to carry weighty additional meaning. This is the case for instance of Li. -drenca, the jan-derivative to drinca ‘to drink’, only attested as gedrenca (demergere) ‘to drown, sink’, as ofgedrenca ‘to drown, sink’ (cp. L demergere) or as underdrenca ‘to throttle’ (cp. L suffocare). This example touches upon two other points which need clarifying with respect to the treatment of data. One is that if the verb is attested only in the past participle, no conclusive information about its valency can be drawn. All formations with -drenca are attested in the past participle. See for instance f. 166rb17 (Luke 10.15) under (6). The Lindisfarne text is from the digitized facsimile of the manuscript in the British Library webpage, cited by folio, page, column and line in that order; the relevant verbs are in bold; all English translations have been adapted from the DouayRheims English version; this one corresponds to the Latin original: (6) et tu capharnaum usque in caelum exaltata usque ad infernum demergeris & ðu þæt is burg oðð heofon ahefen oðð to helle gedrencged ‘and you, Capharnaum, which are exalted to heaven, shall be submerged into hell’. (f. 166rb17; Luke 10.15) L demergeris is in the passive voice and L demergere is a transitive-causative verb, but this need not be the case with Li. gedrenca, attested in the past participle gedrencged in the excerpt above. A past participle, even in a perfect construction with auxiliary verb, can have a transitive reading (he was captured by bandits), an intransitive one (she was gone) and even both (he/the bottle was drunk). Of course, comparison with the use of drencan elsewhere in Old English heavily suggests that it should be transitive-causative in Lindisfarne too, but caution has to be exercised in every case. The other point to make explicit is that the translations of the verbs in this paper, specified in the final list, are contextual; etymological translations are avoided, as far as possible. In the above example Li. gedrenca means roughly ‘to submerge’, even though its etymological translation would be ‘to give to drink’ or ‘to drench’. Similarly, Li. lecga, jancausative to Li. licga ‘to lie’, translates L sternere, substernere only, and contextually means ‘to spread, strew’, not ‘lay’, which would be its etymological translation. Finally, the lemmatization of the attested forms has to be addressed. Present tense forms cannot always be classified as belonging to a strong verb or its weak jan-counterpart. The root vowel should provide information as to whether a particular present form is a reflex of the jan-verb or the strong verb, but given the sound changes affecting particularly 7 diphthongs and long vowels in Old Northumbrian, this is not always the case (Campbell 1959: esp. 110‒112; Brunner 1965: 38‒51, 106‒119). In fact, dictionaries tend to interpret the forms differently. By way of illustration, “(leht fato) beornendo” ‘burn (intr.)’ (f. 134va 2; Luke 7.17; cp. L (lucernas) ardentes) appears under berna (wk) in Cook (1894) and under byrnan (st.) in the DOE. This is a particularly intricate jan-pair, in which even the normalized infinitive forms are disputable. For the purposes of this study, however, the correct ascription of the form beornendo is not decisive, since its meaning and function are found in other forms that clearly belong either to the strong verb or the weak jan-verb. A more significant case is Li. reca ‘to fumigate’, as will be explained in the following section. The data relevant for the study have been collected in the final list attached. The list is the result of both an etymological and a textual study of the verbs involved. The former determines the selection of items and the historical classifications proposed; the latter brings about a new assessment of their meaning and syntactic usage. 4 Data analysis This section presents the results obtained from the analysis of the deverbal jan-verbs and their derivational bases attested in Lindisfarne. Their exact meaning and valency in the gloss will be established, classified and contrasted with those found in other Old English texts. The basic assumption is that those instances where Lindisfarne diverges significantly from other texts point to general directions of linguistic change in Old Northumbrian derivational morphology. In the first subsection (4.1), those Old English jan-verbs unattested in Lindisfarne will be dealt with; those attested will be the concern of the second subsection (4.2) 4.1 Old English jan-causatives unattested in Lindisfarne 35 jan-verbs out of 57 are not attested in Lindisfarne. They are: (7) (a)bylgan ‘to anger, offend’, acwencan ‘to extinguish (fire, lamp); snuff out (a candle)’, ahrȳran ‘to destroy, cause to fall’, astyrfan ‘to cause to die, kill’, aþrȳtan ‘to weary, tire out (so.)’, berȳfan ‘to deprive (so. + acc)’, beswemman ‘to make to swim’, bætan ‘to bridle and saddle; bait (so. + acc/dat)’, cennan ‘to make known, declare’, dwellan ‘to lead into error (so. + acc); err (intr.)’, dyrfan ‘to bring into danger, afflict; engage in (1x)’, flȳgan ‘to put to flight, disperse (so., sth.)’, fyllan ‘to cause to fall, fell, kill’, gremman ‘to enrage, provoke’, leccan ‘to moisten, wet (sth.)’, hnægan ‘to cause to bow; humiliate’, litan ‘to incline (sth.)’, myltan ‘to melt (caus.; intr.); digest’, ræran ‘to cause to rise, rear, raise’, slȳpan ‘to put, slip (sth. + acc)’, scremman ‘to cause to stumble’, sencan ‘to sink (sth.), submerge, drown’, slætan ‘to incite (a beast + acc) in order to cause damage’, smican ‘to emit smoke (intr.); smoke, fumigate (sth.)’, sprengan ‘to scatter; burst (sth.); cause to spring; apply a clyster’, stæþþan ‘to support’, stepan ‘to cause to take a step’, swebban ‘to put to sleep; kill’, swengan ‘to cause to swing; swing, fling, strike’, sȳcan ‘to suckle, give suck’, wyrdan ‘injure, annoy; hinder’, þwænan ‘to reduce the size, cause to dwindle’, þyrran ‘to render dry’, wecgan ‘to move, shake (sth.)’, wyllan ‘to boil (sth.); torment, agitate (so.)’. Most of them are only scantily attested even outside Lindisfarne, but there are two conspicuous absences in the gloss. One is OE myltan, the causative corresponding to the strong verb OE meltan, not attested in Lindisfarne either. The reason probably lies in the lack of context for it in the original text. L fundere, which would most closely correspond to OE 8 myltan, is used in the sense of ‘to pour’ (not ‘to melt’) in Lindisfarne and more accurately glossed with OE ageotan. The other notable absence is OE ræran ‘to cause to rise, rear, raise’, causative to OE risan ‘to rise’, which is attested in Lindisfarne only in the elsewhere rarer sense ‘to be fitting, becoming’ (L debere, licere). For the sense ‘to rise’, the prefixed formation OE arisan is mostly used, glossing L surgere, oriri, ascendere and others. The sense ‘to cause to rise, raise’ is frequent in Lindisfarne, where L tollere is glossed mostly as Li. ahebba. The absence of OE ræran seems therefore to be a genuine gap in the Old Northumbrian lexicon. 8 After studying all the attestations of the relevant verbs in the gloss, one may conclude that none of the jan-verbs in the above list has been functionally replaced by its strong counterpart in Lindisfarne. That is, none of the strong verbs that serve as bases for any of the jan-formations absent in the gloss (see (7)) shows causative meaning in Lindisfarne in addition to its original non-causative sense. This is the case, however, in other Old English texts. As a clear example, consider the strong verb Li. smeca, used only as intransitive ‘to emit smoke’ (L fumigare), whereas OE smeocan can be used both as intransitive ‘to emit smoke’ and transitive ‘to smoke; fumigate (sth.)’. This verb serves as the base to the janformation OE smican ‘to emit smoke (intr.); smoke, fumigate (sth.)’, not attested in Lindisfarne as shown above. Thus, with respect to jan-verbs not attested in Lindisfarne, the preservation of original non-causative meaning and valency in their strong bases in the gloss is a conservative trait concerning linguistic change in comparison with other dialects. 4.2 Old English jan-causatives attested in the Lindisfarne gloss 23 out of 57 secure Old English causatives are attested in Lindisfarne. The meaning and function of 17 of them do not diverge from those in other Old English texts. They are: (8) Li. bega ‘to humiliate’, berna ‘to burn’, gecæla ‘to cool (sth.), cwœlla ‘to destroy’, gedrenca ‘to drown’, græta ‘to greet’, hwerfa ‘to convert’, læda ‘to lead’, læra ‘to teach’, lecga ‘to spread’, generiga ‘to deliver’, reca ‘to smoke’, setta ‘to set’, geswœnca ‘to afflict’, wæcca ‘to watch’, towælta ‘to roll (sth.)’, wœnda ‘to turn’. Little of interest can be gleaned from similarities in a contrastive study, beyond ascertaining their existence, determining the extent of overlap (in this case, roughly three quarters of the verbs show no difference in their usage in Lindisfarne with respect to other varieties) and some linguistic observations, which follow. Li. reca ‘to send forth smoke’ cannot be clearly classified as strong or weak. Its only attestation is the present participle <recende> ‘smoking (flax)’ (cp. (linum) fumigans; f. 49vb5, Matthew 12.20). The form could correspond to (non-Northern) OE reocan ‘to reek, send forth smoke’ with Anglian smoothing, or to the weak verb OE recan ‘to smoke (sth.), fumigate (sth.)’. Both the meaning and intransitive use of Li. reca in this attestation support its interpretation as strong verb. The elsewhere rather common simplex OE nerian is not attested in Lindisfarne. L salvare is glossed with Li. (ge)haeaela, gehalgiga. For L eruere, which in other texts is also sometimes glossed with OE nerian, Lindisfarne uses Li. generiga and genioma. Occasionally, the absence of a verb in a causative opposition can be put in relationship with the existence of a frequently attested synonym. This applies to the strong verb OE 8 For a detailed summary of dialectal differences in vocabulary in Old English see Fernández Cuesta and Rodríguez Ledesma (2001b). 9 cwelan ‘to die’, base to the causative OE cwellan ‘to kill’ (cp. Li. cwœlla). The former is not attested in Lindisfarne, where verbal derivatives of the adjective OE dead (OE deadiga and others) and the verb Li. (ge)swelta are used instead. Finally, Lindisfarne preserves the original meaning of the jan-verb gretan (cp. Li. græta). This is the jan-causative to the strong verb OE gretan ‘to bemoan, weep for’, not attested in Lindisfarne. The usual meaning of the jan-verb OE gretan is ‘to approach, visit, address, greet’. In Lindisfarne it translates L salutare ‘to greet’. The original causative sense of the jan-verb OE gretan is appreciable in the meanings ‘to insult’, ‘to attack’ and the like, which are attested in Old English, but much less frequently than the senses ‘to approach, greet’. Not so in the Lindisfarne text, where they are pervasive in the prefixed formations Li. agræta ‘to throw down’ (cp. L elidere), or even more telling, Li. gegræta ‘to torment’ (cp. L torquere). Again in this instance the Lindisfarne gloss preserves inherited material to a higher degree than other Old English texts. For one of the jan-verbs attested in the gloss, Li. besenca, an alternative reading has been proposed; it will be addressed in more detail below. Five of the 23 causatives attested in Lindisfarne show a different meaning or function from other texts in Old English. They are: (9) Li. depa ‘to dip’, feriga ‘to carry’, læfa ‘to leave’, gescrenca ‘to cause to dry’, stenca ‘to stink’. The diverging instances have varying degrees of significance depending on the robustness of attestation among other factors. They will be dealt with individually. The janverb Li. læfa is attested twice in Lindisfarne (Mark 12.9 and Mark 12.22) with the meaning ‘to leave (sth.)’ (L relinquere), causative to the strong verb OE belifan ‘to be left over, remain’, unattested in the gloss. The jan-verb is attested in other Old English texts mainly with causative sense (‘to leave (sth.)’), but also, though seldom, as non-causative ‘to remain’ (e.g. in Aelfric’s Catholic Homilies, ÆCHom II, 3 21.79, following the DOEC citing system), having incorporated the meaning of the strong verbal base OE belifan. This innovation is not found in the gloss in the simplex. The prefixed formations Li. gelæfa and oferlæfa do have the intransitive sense ‘to remain’, glossing L manere and superesse, respectively; on Li. gelæfa, see below. Two other jan-verbs remain faithful to their original causative meaning in Lindisfarne, namely Li. depa and gescrenca. The first is attested only once (Matthew 23.26), translating L intingere ‘to dip, lower into or immerse in liquid’, causative to a strong verb OE dufan ‘to dive, plunge, sink (intr.)’, unattested in the gloss. 9 The jan-verb has developed the specialized metaphorical meaning ‘to baptize’ outside Lindisfarne, where it is well-attested (normalised as OE dypan) in this sense. In the gloss, however, L baptizare is glossed as Li. clænsiga, fulwiga, gefulwiga and ingefulwiga, never as Li. depa. As for Li. gescrenca ‘to dry (sth.)’ (strong base Li. gescrinca ‘to wither away, dry up’), it is only in Lindisfarne that the straightforward causative meaning appears. Its only occurrence follows, with the English translation: (10) sole autem orto aestuauerunt et quia non habebant radicem aruerunt sunna uutedlice miððy arras weron forbernedł besenced & forðon ne hæfdon ł næbbend 9 For the ultimately unexplained sound relationships in this verbal root see Seebold (1970: 155‒156), Heidermanns (1993: 153‒154) and García García (2005: 102‒103). 16 Appendix Deverbal jan-verbs with possible causative meaning attested in the Lindisfarne gloss The following alphabetical list includes all potentially causative jan-pairs attested in Lindisfarne and their function in the Lindisfarne gloss (Li.), general Old English (OE; see n. 2) and Germanic (PGmc). In the first column of the table the jan-verb is given in its (conjectured) Lindisfarne infinitive form with Present-Day English translation and Latin original in italics. Below it, in the same line, the normalized Old English form is listed with its meaning (revised from García García 2012a: 143‒148). In the second column, the corresponding strong verbal bases in Lindisfarne (if attested) and Old English are listed as before. The third, fourth and fifth columns label the relationship between jan-verb and base in Li., OE and PGmc. The following labels describe the function of the jan-formation or the semantic relationship between jan-verb and strong base, as the case may be: C = causative Col C = collapsing causative opposition, with one or both members adopting new valency values (see section 1) IDI = a Germanic causative relationship has undergone idiosyncratic semantic changes SPE = semantic specialization in either member of the jan-opposition. JAN-VERB BASE FUNC. IN LI. FUNC. IN OE FUNC. IN PGmc Li. bega ‘to humiliate’ (trans.); ‘bend (the knees)’ (only past part) (cp. L humiliare, flectere) OE bigan ‘to bend (trans.; intr. (-)); turn (caus.; intr.)’, ‘humiliate’ Li. gebuga ‘to bow, bend; cut down’ (trans.) (cp. L caedere, inclinare) OE bugan ‘to bow, bend (intr.; trans.?); turn (intr.)’ Col C+IDI Col C+IDI C Li. berna ‘to burn, light’ (trans. and intr.) (cp. L accendere, ardere, comburere) OE bærnan ‘to burn’ (trans.; intr. (-)) Li. bearna ‘to burn’ (intr.) (cp. L ardere) OE byrnan ‘to burn (intr.; trans.(-))’ Col C Col C C Li. gecœla ‘to cool (sth.)’ (cp. L refrigerare) OE celan ‘to cool or chill (sth.), make cold; quench (thirst) Not in Li. OE calan ‘to be or become cold; to make cold’ C Col C C Li. cwœlla ‘to destroy, kill’ (cp. L interficere) OE cwellan ‘to kill’ Not in Li. OE cwelan ‘to die’ C C C Li. depa ‘to dip, lower into, or immerse in liquid’ (trans.) (cp. L intingere) OE dypan ‘to dip, immerse in liquid (sth.); baptize’ Not in Li. OE dufan ‘to dive, plunge, sink (intr.)’19 C C+SPE C Li. gedrenca ‘to drown, sink’, trans.? (only ge- + past part) (cp. L demergere) Li. drinca ‘to drink’ (cp. L bibere) C C C 19 To p/f variation in this verb see Seebold (1970: 155‒156). 17 OE drencan ‘to give drink to; drench, saturate’ OE drincan ‘to drink, drink’(sth. + acc) Li. feriga ‘to carry; go, pass’ (cp. L ferre) OE ferian ‘to carry, transport’ Li. fara ‘to go travel’ (cp. L ire, etc.) OE faran ‘to go, travel’ Col C C SPE (by ship) Li. grœta ‘to greet’ (cp. L salutare); Li. gegrœta ‘to torment’ (cp. L torquere) OE gretan ‘to approach, touch; damage, attack; address (so. + acc); greet’ Not in Li. OE gretan, greotan ‘to bemoan, weep for’ C+IDI C+IDI C Li. hwerfa ‘to convert (intr.); borrow’ (cp. L convertere, mutuari) OE hwyrfan ‘to go, return; turn, change (caus.; intr.); exchange’ Not in Li. OE hweorfan ‘to go; turn, change (intr.; caus. (-))’ Col C Col C C Li. læda ‘to lead, carry’ (cp. L ducere, adducere, educere, tollere, ferre, conferre) OE lædan ‘to lead, take, carry, bring, produce’ Not in Li. OE liþan ‘to go, sail’ C C C Li. læfa’ leave (sth.)’ relinquere OE læfan ‘to leave; remain (-)’ Not in Li. OE belifan ‘to be left over, remain’ C Col C C Li. læra ‘to teach, instruct’ (cp. L docere, instruere, admonere) OE læran ‘to teach; preach; persuade, suggest’ not in OE; Go. lais ‘knows’ C C C Li. lecga ‘to spread, strew (sth.)’ (cp. L sternere, substernere) OE lecgan ‘to cause to lie, lay; slay’ Li. licga ‘to lie, be at rest, lie down’ (cp. L jacere, discumbere) OE licgan ‘to lie, be at rest; lie dead’ C C C Li. generiga ‘to deliver, pluck out’ (cp. L eruere) OE nerian ‘to save’ Not in Li. OE nesan ‘to be saved from, escape from’ C+IDI C C Li. reca ‘to smoke’ (intr.) (cp. L fumigare) OE recan ‘to smoke (sth.), fumigate’ Not in Li.?20 OE reocan ‘to reek, send forth smoke’ ? C C Li. gescrenca ‘to cause to dry’ (cp. L arere) OE screncan ‘to cause to stumble, ensnare’ Li. gescrinca ‘to wither away, dry up’ OE scrincan ‘to wither away, dry up; become weak; shrink’ C C+IDI? C? Li. besenca ‘to scorch’ (cp. L aestuare) OE sengan ‘to singe, burn slightly; afflict’ Li. singa ‘to sing’ (cp. cantare, canere) OE singan ‘to sing’ C C C Li. setta ‘to set, place, put, settle’ (cp. L imponere, ponere, statuere, instituere, constituere) Li. sitta ‘to sit, be seated, sit down’ (cp. L sedere, discumbere) C C C 20 The occurrence in Lindisfarne might be an attestation of the strong verb OE reocan (see section 4). 18 OE settan ‘to set, place, put; settle, subside (intr.) (-)’ OE sittan ‘to sit, be seated; occupy (a seat)’ Li. stenca ‘to stink’ (cp. foetere) OE stencan ‘to scatter; emit breath with effort; stink’ Not in Li. OE stincan ‘to spring, leap; emit a smell’ SPE Col C+SPE C Li. geswœnca ‘to afflict’ (cp. L vexare) OE swencan ‘to cause a person to labour, harass, afflict’ Not in Li. OE swincan ‘to toil, labour, work with effort’ C C C Li. wæcca ‘to watch, be awake; raise, provoke’ (cp. L vigilare, suscitare) OE weccan ‘to waken, arise, spring (intr.; caus. (-))’ Not in Li. OE wæcnan ‘to come into being, be born, spring’ Col C Col C C Li. to/ge/a/efta/fromawælta ‘to roll (sth.)’ (cp. L advolvere, revolvere) OE wyltan ‘roll (sth.)’ Not attested in OE; OIc velta ‘to roll (intr.)’ C C C Li. wœnda ‘to turn’ (intran.) (cp. L verti, reverti) OE wendan ‘to turn (round), change (caus.; intr.); go (refl.; intr.)’ Li. winda ‘to plat’ (cp. L plectere) OE windan ‘to spring (intr.); roll (intr.; caus.); weave (sth.)’ Col C+SPE Col C+SPE C