Layout

The converb, in its least specific and sharp resolution, is The term ›converb‹ was coined, so far as I know, used to mean ›adverbial verb form‹, or ›verbal adverb‹ without definition, by the Finnish Altaicist G. Ramstedt (see the subtitle of HASPELMATH & KÖNIG 1995). Mostly as ›Converbum‹ or ›Converbium‹, in his 1902 mono- and for long it has been known, in the description of var- graphic description of Khalkha, the most famous of ious languages, as ›gerund‹1. Definitions of the converb Mongolian dialects5, which is the basis of today’s stan- reveal an underlying blurredness: HASPELMATH (1995: dard Mongolian. The term is still widely used in Altai 3 ff.): ›Non-finite verb-form whose main function is to and Turkic linguistics, where it is considered by its users mark adverbial subordination‹2; NEDJALKOV’S (1995) is to be apt, referring to a notion ›not existing in Indo-Eu- more sophisticated: ›a verb-form which depends syn- ropean‹. Following Ramstedt, the term was used by tactically on another verb-form but is not its syntacticactant, that is does not realize its semantic valences‹.
(This is surely unsatisfactory, for the converb is ar- The general paragraphs which form the first part of this paper guably actantial in cases like ›start walking‹). Probably overlap and to a degree repeat the general introduction to a paper the worst is the definition in HIMMELMANN & SCHULTZE- on converbs in Egyptian and Coptic by the present writer, present - BERNDT (2005: 60): ›we use the term converb for »par- ed in 2006 at the NACAL 35 conference in San Antonio, Texas ticiples« which are used primarily as adjuncts‹. As (SHISHA-HALEVY forthcoming). Terms like ›adverbial participles‹ or ›gerund‹ are to my knowledge not currently used in Celtic lin- RØNBECH (1979: 35) says of Turkic postpositions and guistics for this notion (consider Irish gerind, geireann, later Welsh gerundial forms, the converbs are ›fluid and hard to hold gerwnd, gerwndin, all for the Latin inflecting verbal noun, but also on to‹, which, for a ›cross-linguistically valid category‹ the Lat. participles and -tum supine).
(HASPELMATH & KÖNIG 1995, in which see Haspel- 01 Historically a misnomer, for some reason especially widespread math’s and König’s own contributions), is not an ideal in English-language writing, more or less corresponding to theFrench ›gérondif‹ (English ›gerundive‹ must be a gallicism). See condition. And indeed, one detects symptoms of termi- GOLDENBERG (1977: 489–99, 2002: 28–30). LEWIS (1967: 174) nological and descriptive insecurity or malaise in the explains his preference of ›gerund‹ over ›converb‹ by the ›merit distinction between ›canonical‹ and ›non-canonical‹ of brevity‹ (!); see also his Chapter XI on ›deverbal adverbs‹.
converbs; between ›general‹ and ›adverbial‹ converbs 02 ›Embedded/incorporated to the superordinate clause‹ (HASPEL- (HASPELMATH 2004: 232 ff.), or ›contextual‹ as against MATH 1995: 8) is no less question-begging.
03 KORTMANN (1995: 196 ff.), on cases like ›Our children hate John ›specialized‹ converbs; symptomatic is also the use of quotes, or of prefixes such as ›pseudo-‹ or ›old-‹ or 04 If the Celtic converbal systems seem less obscure, ›simpler‹ than ›half-‹ (consider ›Old Perfective‹, ›Parfait ancien‹, the English ones, it is only because and in the sense that there is ›Pseudopartizip‹, all for the Egyptian Stative, an emi- in Celtic less homonymy – diachronically, merging, since (for in- nent converb; also ›half-gerunds‹, or ›half-participles‹ stance) the prepositional prefixes mark and distinguish individ-ual, mutually opposed converbs; still, the systems are as complex in Baltic grammatical terminology),3 along with certain red herrings which I find no less than pseudo-queries, 05 RAMSTEDT (1902: 55): ›Da sie (i.e. the Converbs) aber im Khalk- e.g. of polysemy vs. vagueness (KÖNIG 1995: 59 ff.).
hassischen eine wichtige Rolle spielen, habe ich die The issue of the English -ing – converb – ›gerund‹, sagenden und in viel engerer Bedeutung angewandten NamenSupinum und Gerundium durch die hoffentlich deutlichere, ge- ›adverbial‹ and non-adverbial participle, verbal noun eignetere Benennung »Converbum« ersetzt.‹ See also pp. 44 ff., and infinitive, the notorious terminological-conceptual 61 ff., 104 ff. etc. Ramstedt, always historically and morphologi- muddle involving -ing forms and constructions in cally associating the Mongolian Converbs with verbal nouns, English – is a fine illustration of the havoc wrought by still uses here ›Gerundium‹ (e.g. 76 f.), despite his explicit rejec- simplistic obstinate superimposing of a ›prefabricated‹ tion of this term as ›meaningless‹. In the poshumous Russian ver-sion of Ramstedt’s Introduction to Altai Llinguistics (Moscow primitive, essentially morphological model on a com- 1957), the term used is gerundivno-priçastnaya forma (e.g. 111): plicated reality of syntactic dynamicity and sophistica- deepriçastiya is, of course, the Russian correspondent of ›ge- Poppe 1951/52, by A. von Gabain, by Menges (con- word-class aspect is, structurally speaking, more impor- verbs and participles constitute the ›nomina verbalia tant than any other). Where and what is the adverb? Can temporum‹, used alongside ›Gerundium‹ and ›gerundial ›adjunctality‹ represent this category at all? And is this forms‹, also ›Verbal Adverbs‹; see also RAHMATI 1928, quality, thus conceived, of any importance for under- GRØNBECH 1979, AALTO 1987: 186) In 1951, the term standing the converb, when we consider such an over- was applied by Hans-Jakob Polotsky to Ethiopian lan- ruling set of parameters as syntactical slotting? Adver- guages, notably Gurage and Amharic (POLOTSKY 1951: biality, if we stay rigorously analytical (a ›word-class‹ 41 f., GOLDENBERG 1977: 489–94)6, as a syntactical cat- distinct, in distinct paradigmatic commutation, in dis- egory, ›converb‹ specialized, co-existing with, and con- tinct syntagmatic slotting), is too fragmented to be use- trasted to, the morphological ›gerund‹ (POLOTSKY 1951: ful even as an overall umbrella. Adjunctal; adnominal 45, GOLDENBERG 1977: 491). In 1995, the converb was (especially instructive, for often opposed as adnexal to celebrated as a typologically important ›Cross-Linguis- the attributive relative); adlexemic (valential and non- tically Valid Category‹ by Martin Haspelmath and valential); ad-(verbal)-nexus; adclausal; rhematic; rhe- Ekkehard König, editing an anthology of studies that matic (›predicative‹) complement8: adnexal; focal; top- point to the importance of converbal forms on Finnish, Slavonic, Japanese, Asian and European languages of Moreover, there exist (in Celtic as in other lan- different genetic affiliations, For some reason, Ethiopian guages) formal statuses in which adverbiality cannot be is not represented beyond a single footnote;7 neither is recognized and resolved as such, or seems irrelevant; Celtic in Europe, nor Egyptian in Africa. However, it for instance, as formally differentiated from substan - appears the converb is ubiquitous – converb-less lan- tivity or nominality (preeminently in topic or rheme guages seem to be the exception, not the rule, and may status). The immediate or conventional adjunctal asso- be of interest on that very account.
ciation of adverbiality (e.g. in NEDIALKOV 1995: 98; Iquote, ›an adverbial in a simple sentence‹), does notconnect ohne weiteres with the other two alleged con- verbal roles (ibid.), namely ›secondary or coordinatepredicate‹ and ›predicate of a subordinate clause‹.
Problematik, methodological and theoretical issues:Following a cluster of terminological-conceptual theo-retical reflections, I wish to no more than hint here at a systematic consideration of Welsh and Irish verbal ad-verbial-slot and adverbial-commutation features, in this Is the converb in essence and by definition a non-finite scope merely isolating them and appreciating their in- or a finite verb form, or either? Is this essentially mor- ventory and distribution – here not evaluating their phological distinction at all important, especially since systemic standing and their structural profile. The ben- the infinitive, and indeed the converb, may be finitized efits of terminological-conceptual deliberation, whether by various constructional devices. Moreover, the infini- comparative-contrastive or internal-typological, are ob- tive or participle constituents of a Nominal-Sentence vious. There is, I believe, an exercise no more salubri- type nexal pattern, are ›finite‹ in construction and inter- ous than judging the degree of comparability of linguis- dependence with their theme or subject. The typology tic phenomena, putting in sharper focus and forcing us of actor-expression of converbs as compared with in- to contemplate critically notions we take for granted and use automatically, almost thoughtlessly. (The argument,often explicitly or implicitly advanced, that terminologyis a trivial concern is unacceptable: not only do terms,by evoking concepts and conceptualization, subtly, in- 06 See MYHILL & HIBIYA (1988: 355 ff.), on the converb as a narra- sidiously guide our descriptive view and insight; felici- tive form in the Gurage language Soddo.
tous revisionist terminology may suggest (cor)relations 07 HASPELMATH & KÖNIG (1995: 342 n. 52): ›Thus, Amharic, which that would otherwise be invisible or hidden, or ignored is not genetically related to Turkic, and has had no close contactswith it, exhibits a very similar constituent order and correspon- ding patterns of converb subordination.‹ POLOTSKY (1951) is notquoted; nor is POLOTSKY (1965), typologically comparingAmharic and Turkish syntax.
08 In verbal ›Secondary Predication‹ syntax, converbs play a strik- ing role: see HIMMELMANN & SCHULTZE-BERNDT (2005).
09 Cf., for Modern Welsh and Modern Irish, SHISHA-HALEVY (1998: First, we must contemplate the descriptive meaning of 56 f.); o- finitizes narrative infinitives in Welsh, especially Mid- ›adverbiality‹, refining our conception of this most dif- dle Welsh, do- narrative converbs in Irish; see also SHISHA- ficult, and perhaps most dubious of word-classes (the HALEVY (1998: 264, s.v. ›i- cum infinitivo‹).
An essentially junctural set of parameters concerns was singing that I was‹; another case of converb/infini- the converb as a unit, continuous or discontinuous. An- tive homonymy is, I believe, in narrative-carrier slot alyticity and syntheticity are observable as both dia - chronic and synchronic qualities of converbal forms.
As an aspect of juncture and a symptom of advanced Almost a curiosity in this context is the question of grammaticalization,10 the synchronic absence in Celtic unity, which is largely psychological, namely a bias (sometimes irregularity) of infinitive mutation inside against viewing the [prefixed preposition+noun] syn- some converbs is striking (Irish ag-, Welsh yn-, wedi- ; tagm – so in Celtic – as a single converbal unity, which this is only superficially paradoxical, for this feature in- is clearly a grammaticalized ›morphologized‹ pattern; dicates reduced analyzability (conditioned mutations, apparently, there’s no such difficulty with postpositions beside marking rection and rectum status, promote ana- and case endings. Another formal question is one of ex- ternal juncture: converbs, especially non-finite ones,are often ›induceable‹ for verb categories such as tense,mode, person or negation. The scale of this induction is a significant factor of linkage and delimitation about theconverb.
The converb’s structural identity, as defined by commu-tation, compatibilities, relationships (synchronic or di-achronic), distribution and affinities with verbal nouns (notably infinitives), not merely morphologically (theAltai Converbs are synchronic or diachronic case-forms ›Syntaxic‹ features: ›Ordination‹. A question recurring of verbal nouns), but systemically, in the sense of the in general and specific accounts of Altai and Turkic con- implicative significance of their respective performance verbs: what is the ›main‹ and what the ›subordinate‹ (Leistung) and their mutual trade-off, homonymy (where verb, or action, or predication, in cases corresponding to applicable) and/or complementary distribution.11 ›he fell slipping‹, and especially to ›he started crying‹,›he goes on crying‹ – descriptive or auxiliary or modalverbs, typical of Altai and Turkic languages but impor- tant elsewhere. The semantic query may perhaps be dismissed as subjective and non-illuminating, and it of- Once again, we face the theoretical conundrum of the ten appears to be trivial, leading to such problematic epistemology and phenomenology of ›ordination‹, or distinctions such as ›logical‹ as against ›grammatical‹ ›inordination‹ (HAMP 1973: cf. Einordnung): the rela- ›main-ness‹. Ramstedt himself, trying to make sense of tion, and indeed the macrosyntactic patterning of one the elusive ›Haupthandlung‹ and ›Nebenhandlung‹ (not predicative nexus adjoining another, that is neither sub- Hauptsatz and Nebensatz) hierarchy, has recourse to ordination, nor coordination (cf. KAZENIN & TESTELETS the not really helpful ›psychologisches Hauptwort‹ (as 2004): the adnexal expansion of one nexus by another, against ›grammatical‹ or ›syntactical‹?), for the converb and of a noun syntagm or pronoun by a nexus (the latter opposed in-paradigm to attributive expansion). ›Subor-dination‹, which is a basic component of prevailing con-verb description, is, I believe, a faulted concept in syn- tactic analysis, and at the very least is question-begging,and not merely in ›esoteric‹ languages (including Grammaticalization: Grammaticalization and analyz- ability are kindred, correlated properties, both a matter Syntactic versatility and environmental sensitivity of of gradience, and a function of diachrony – the earlier the converb: privilege of occurrence as adjunctal ex- occurring, the less sharply analyzable a syntagm. More- pansion (›She turned to me dancing‹), adnominal ex- over, as noted above, cases of preposition governing in- pansion (›The girl dancing was too lovely for words‹), finitives, preeminent candidates for converb-hood, are, adnexal expansion (›I found her dancing‹), rheme (›She I suggest, not synchronically analyzable as preposi-tional phrases but grammaticalized as converbs. Infini-tives ›alone‹, too, as narrative carriers – whether astightly-linked verb-serialization forms, or (for instance) 10 For paradigmatic-junctural correlates of grammaticalization, see as dialogue response-forms – are arguably adverbial SHISHA-HALEVY (2003b, 2004): the prefixes of Irish and Welsh›infinitival converbs‹ are, arguably, not synchronic prepositions and thus converbal. Note too the fact that the focalized at all, also by token of their paradigms.
converb in ModW, the converb as Cleft Sentence focus, 11 See HASPELMATH (1995: 28), ›a kind of infinitive‹ (KORTMANN is homonymous with the infinitive: Canu yr oeddwn ›It was dancing when I first saw her‹), focalizability (›It written as a continuous junctural unit.) In brief, Welsh was dancing that I first saw her‹), topicalizability and Irish converbs are formed by prefixing preposition- (›Dancing, the girl looked at me‹), and others, e.g. ex- homonyms to the infinitives, or are infinitive-homony- mous; in both cases, converbs may be finitized.
Textemic significance and properties of converbs, espe- In Welsh, this is a richer paradigm (category) of infini- cially narrative concatenation by converb sequencing, tive-base converbs than the rhematic one, with their pre- often discussed in HASPELMATH & KÖNIG (1995).
cise individual semantics still unclear.14 (1) Tawodd, g a n - w l y c h u ’i wefusau. ›He fell silent, (2) Bwytai’n araf d a n - e d r y c h o’i chwmpas. ›She ›Adverbiality‹ yet again: what does this actually mean, (3) Cododd yntau d a n - g r y n u . ›He arose, trem- functionally speaking, in an analytical view of lan- guage, one that rejects the Part of Speech model as apri- (4) Ni ddywedodd yntau ddim w e d i - e i - d d a r l l e n oristic and logic-based? As a structurally conceived … ›He did not say anything after reading it.‹ word-class, the adverb is not a category at all, but a syn- (5) Oedd, mi ’roedd o yno, y n - s i a r a d ac y n - thetic and synthesizing cluster of numerous paradigms.
l o l i a n efo thair o enethod … ›Indeed, he was there, talking and fooling around with three girls.‹ (6) Gweld ei fam y n - g o r f o d gweithio’n galed ag yntau ’n - e n n i l dim … ›Seeing his mother having to (7) A - b a r n u oddiwrth ei olwg, nid oedd dim yn de- The converb an adnexal satellite (expansion-form). As byg i siopwr ynddo, yn ôl ei barn hi. ›To judge from his brilliantly pointed out by Johannes LOHMANN (1965: looks, there was nothing in him resembling a shop- 295, see SHISHA-HALEVY 2009), the converb has strik- ing affinities with the original and historically ›correct‹ (8) Nid adwaenai ef, a g - e i t h r i o ar brynhawn Sad- conception of the participle – μετοχή, not ›partaking of wrn a dydd Sul, ond fel le i chwi ddychwelyd ar ôl di- the nature of verb and noun‹ or similar, but ›predication- wrnod o waith … ›He did not know it, except for Satur- sharing‹ or ›in predicative union‹ (other more or less day afternoon and Sunday, but as a place for you to re- descriptive appellations, by Lohmann and others, are ›durchkonstruierter Satz‹, ›adsentenziale Subordina- (9) Crychodd ei thalcen w r t h - g o f i o am Idris … tion‹. The intriguing status and slot that has been vari- Anghofiodd ef w r t h - g l y w e d tramp y traed ar yr heol ously called ›halb-prädikativ‹ (Behagel), ›degradiertes galed. ›She frowned, upon remembering Idris … She Prädikat‹ (Hermann Paul), ›Prädikative Apposition‹ forgot him upon hearing the tramp of feet on the hard (Sommer), ›copredicative‹ (HASPELMATH 1995: 17 ff., of participles – nearest to Lohmann’s ›predicative shar- (10) Bhíos i-mo-shuí sa doras, a g - c u r caoi ar líonta ing‹); ›second‹ or ›secondary‹ predicate‹ (HIMMELMANN ronnach. ›I was sitting at the door, repairing mackerel & SCHULTZE-BERNDT 2005), all of which I prefer to call, following Otto Jespersen, ›adnexal‹ (SHISHA-HALEVY2007: 695 s.v.). The relationship of converbiality withrhematicity, as a special sector of the clause-linkagespectrum, is here at issue.12 On a yet higher general 12 The current concept of ›participant orientation‹ (e.g. in HIMMEL- plane, I would pose the question of the affinity of ad- MANN & SCHULTZE-BERNDT 2005) is germane here. This is es-sentially a complicated junctural feature, of the two predications verbiality to predicativity or rhematicity, an affinity inter-merging in looser linkage with common actants.
manifested by formal similarities – consider the Arabic 13 Examples are quoted from fiction by (ModW) Kate Roberts, accusative, Welsh lenition, or Coptic n-marking.
Saunders Lewis, Islwyn Ffowc Elis, T. Rowland Hughes; In the latter part of this paper, I shall suggest and (ModIr.) Myles na gCopaleen, Pádraic Breathnach, Pádraic Ó illustrate13 Irish and Welsh candidates for converb- Conaire, Liam Ó Flaithearta. The examples are selective and rep-resentative.
hood, as well as syntagms or elements of relevance to 14 The present writer is engaged in a special study of the construc- the study of converbs. in a brief commented typology.
tion and meaning of converbs in narrative, as part of a compre- (Note that all converbs are emphasized; converbs are hensive study of Kate Roberts’s narrative syntax.
(11) D’airigh sé go raibh sí amuigh a g - d a m h s a . (18) Yr ydych w e d i - g o r f f e n nofel, meddwch. ›You ›He noticed she was outside dancing‹.
(19) Nid wyf w e d i - p r y n u llyfrau Cymraeg ers tro. ›I haven’t bought Welsh books for a while.‹ (20) Mae M. y n - e i - d d a r l l e n rŵan; ac yr wyf w e d i - e i - d y n g e d u nad yw i - s i a r a d efo mi. ›M. is The converb completing a descriptive action-phase- reading it now; and I have condemned him not to speak marking finite verb: In information-structure terms, the converb is here a rhematic (predicative) complement.
(21) Ydach chi wir-yr a m - b r y n u car? ›Are you (12) … gweld defaid a gwartheg y n - p o r i ’n hapus a hamddenol … ›… seeing sheep and cattle grazing hap- (22) Cawell wyt ti i - f o d i - d d w e u d . ›It’s »cage« (13) Stop an stócach óg a g - c a i n t . ›The youth (23) Dydach chi ddim a m i - a d a e l o … yn nag ydach, Mari? ›You are not going to leave him … are (14) Tosnaigh s é a g - r i t h . ›He started running.‹ (15) D’éiríos i - m o - s h e a s a m h . ›I rose up (lit. arose (24) Pwy sy’n - d w e u d fy mod i a m - d y n n u ’r drol? ›Who says I’m going to pull the cart?‹ (16) ›Tá mé tinn tuirseach a g - b re a t h n ú ort is a g - (25) Tá siosarnach na bhfeithidi a r- s i ú l fós. ›The é i s t e a c h t leat!‹ is chuir sí ina-sheasamh amuigh le balla sa gcúinne é. ›»I’m sick and tired watching you (26) Bhí a dhá suil ar-lasadh. ›His two eyes were and listening to you!« and she made him stand (lit. put him standing) outside against the wall in the corner.‹ (27) Tá tú l e - p ó s a d h ! ›You are about to get mar- (17) Dhearbhaigh sé dhi, lom díreach, go raibh sé a g - d u l a g - p ó s a d h go luath. ›He declared to her, (28) Bhíos i - m o - s h u í sa doras, a g - c u r caoi ar straight out, that he was going to get (lit. getting) mar- líonta ronnach. ›I was sitting at the door, repairing (29) Tá sí b á i t e . ›She has/is drowned.‹(30) Tá sé r á i t e go … ›It is said that…‹ (31) Cad a bhí d é a n t a agam? ›What had I done?‹ (32) Ní raibh an tae ó l t a agam … ›I had not (yet) The converb supplying periphrastic durative-statal andperfectal components of the verb-system.
(a) Observe in Welsh the typical interlocutive envi- ronment of the rhematic wedi-converb.
Only for non-specific substantival nucleus? (b) The ›future‹ converbs in Welsh: deontic i-, ›tem- (33) … swn cath y n - c e rd d e d ar garped. ›… the pus instans‹ am-: none expresses a purely temporal fu- (34) … fel carcharor w e d i ’ i - d d e d f r y d u i far- (c) The embracing nexal negation of this pattern in wolaeth. ›like a prisoner condemned to death.‹ Welsh, by means of ni-… (ddim) must be kept distinctfrom the negation of the constituental converb itself (forwhich see below).
(d) Observe in Irish the almost full resemblance of the intransitive converb to the incidental noun predica- In Modern Welsh, the yn-converb in focal status is tion (see SHISHA-HALEVY 1998: 193 ff., 201): Bhí Fear- homonymous with the infinitive, but different in the danand ina sheanduine críonna. ›F. was a wise old man.‹ topical construction of the Cleft Sentence: y-conver- (e) In the same rhematic slot in this predication pat- sion, not a-. The converb may be part of ›Envelope Fo- tern we find in Irish also ›ppp‹-type (-ta/te) participles, cussing‹ (SHISHA-HALEVY 1998: 28, 31).
which are here considered and illustrated as converbs.
(35) C l y w e d yr oeddech chi, g w e l d oedden ni. However, adjectives that share this slot (tá sí óg ›she is ›You were hearing; we were seeing.‹ young‹), while still ›adverbial‹ in the paradigmatic (36) Nid g w a r a f u n pleser i’w thad yr oedd, ond sense, cannot be converbal (except in the sense of ›rhe- g w a r a f u n caethiwed ar ei threfniadau ei hun. ›She matic adverbials‹; ›non-verbal converbs‹ is a contradic- was not refusing pleasure to her father, bur refusing en- tion in terms; SHISHA-HALEVY 1998: 202 ff.).
(37) G w e i d d i ormod yr ydan ni. ›We’re shouting am y ddwy ffured. ›Then W. arising, and going to the (38) Tr i o osgoi’r cwestiwn yr ydach chi, yntê? (50) Ann y n - g w y l l t i o ac y n - g w e i d d i digon i bobl ›You’re trying to evade the question, aren’t you?‹ y draws nesaf ei chlywed … Sam a Bob, y ddau filgi ’n - (39) I n a - s h e a s a m h agus a dhá lámh ina phócaí e i s t e d d un o bobtu Wmffra y n - e d r y c h dan eu cuwch aige ag breathnú ar bhád beag béal fúithí ar an duirling arno … ›Ann flying into a passion and screaming a bhí F. Mac F. Lá Fhéile Sin Seáin … ›Standing on the enough for the people next door to hear her … Sam and beach with his two hands in his pockets, looking at a Bob, the two greyhounds, sitting on each side of W.
small boat with an opening in her bottom, was F. Mac F.
(40) A g - m a g a d h fum a bhi sé, a g - m a g a d h faoin amadán a chaith a chuid leis. ›He was mocking me, mocking the fool who spent his substance on him.‹ (41) Níd a g - é i s t e a c h t leis atáim, ach go bhfuil sé The difference and interplay between the two converb a g - d u l tríom. ›I’m not listening to him: he is rather go- forms is subtle and not easy to describe. In autobio- graphical diary-style or logbook-style narrative, yn-converbs are used for locutive (1st-person) pivotalevents, and/or for durative action (SHISHA-HALEVY (51) G w e i d d i ar Meri’r eneth hynaf i’r ty wedyn, (42) Wr t h - w e l e d 15 y cymylau duon a grogai fel iddi redeg ar neges i’r siop … ›Shouting for M. the old- bwganod dros y môr troes ei wyneb at Gwm Dugoed eil- est daughter to the house, for her to run to the shop on waith. ›Seeing the dark clouds which were hanging like bogeys over the sea, he turned his face towards Cwm (52) C y c h w y n d a n - c h w i b a n u ’n hapus … ›Set- (43) Wr t h - e d r y c h ar y llawr, yr oedd yno llanast (53) Diwrnod golchi. We d i - rh o i fy nghas ar y anghyffredin. ›Looking at the floor, there was an un- peiriant golchi … ›Washing day. Put (lit. having-put) my pillow-case on the washing machine …‹ (44) We d i - g a d a e l y Bont a gadael llawer o bobl ar (54) We d i - d i f f o d d y goleuadau i gyd ond yr un ôl yno, lledodd y gorwelion … ›After leaving the Bridge dwaetha, c o d i fy ngolygon at y Mona Lisa, d i m - y n - and leaving many people back there, the horizons h o f f i ei golwg o gwbl. ›Put (lit. having-put) out all the lights but the last, raising my eyes to the Mona Lisa, not (45) Ach a g - i a r r a i d h dul suas an cnoc d h ó , chuala sé caoineadh caol crua … ›But, attempting (lit.
(55) Y gath y n - n e i d i o ar fy nglin; y n - e i - h e l i he-attempting) to climb up the hill, he heard shrill, hard lawr. P e n d e r f y n u peidio â mynd i’r capel yfory. ›The cat jumping onto my knee; chasing her down. Decidingnot to go to Chapel tomorrow.‹ (46) Wedyn, dyma g y c h w y n i’r cae … ›Afterwards, a- + infinitive to complete a hyper-event (SHISHA- they set out (lit. approx. here is setting-out) to the field (56) Diolchodd Harri a - c h u s a n u Greta. ›Harri (47) Dyna hi w e d i - d w a d . Approx. ›Here she is (having) come.‹ (›la voilà venue‹) (57) Cododd Catrin a - m y n d i eistedd. ›Catrin rose (48) Sin é anois é a g - d a m h s a agus a g - c e - o l t e o i re a c h t . Approx. ›Here he is now, dancing and (58) Aeth allan a’ i - g a d a e l . ›He went out and left 15 In this slot, wrth represents the entire paradigm wrth-, gan-, 16 Cf. SHISHA-HALEVY (1998: 184 ff., and 1999): App. ii, for the (49) Wedyn Wmffra’n - c o d i , ac y n - m y n d i’r cefn converb in narrative presented nexus in Middle Welsh.
Converbs with infixed lexemic modification (59) M e d d w l am ddoe, a d e c h r a u poeni … Mar- A striking, difficult construction: lenition (if any) of the giad y n - g a l w. D w e u d yr hanes wrthi … ›Thinking of lexeme is a signal of its relation with the ›prepositional‹ yesterday, and starting to worry … M. calling. Telling prefix, not governed by the prelexemic modifier, which appears in a ›phantom slot‹ in the syntagm.
(60) Poli w e d i - g y r r u Meg yn wallgof yn y gegin. (73) Yr oedd Rhys y n - h a n n e r- c y s g u w r t h - g a u ›Poli drove (lit. having-driven) Meg crazy in the ei ddillad. ›Rh. was half-asleep upon buttoning his (74) Yr oedd w e d i - d i rg e l - g re d u yr edifarasai Aels am yr hyn a wnaeth. ›He secretly believed Aels (75) We d i - l l a w n - d d e f ro Probably most challenging for the converbal interpreta- lawenydd mai dyma’r dadmer. ›After fully waking-up, tion of verbal forms, as well as instructive for the ag- he realized, to his joy, that that was the thaw.‹ glutinative typology of Irish and Welsh, is the specificmorphological marking of a substantive or pronoun asthe theme-agens constituent of a finite converb, or of a pronominal object actant of the converb’s lexemicrheme-constituent. Welsh object and agens actants are The question of Celtic converbs negatived is difficult.
infixed, the former by homonyms of the possessive ar- Welsh heb-, elsewhere ›without‹, inflects like any ticles, the latter i-marked (as the ›i- cum infinitivo‹ that- preposition. It appears to negative the converb as ad- form;17 SHISHA-HALEVY 1998: 264 s.v.). In Irish, too, verb, neutralizing yn-, wrth-, gan- and wedi- converbs possessive-article-like infixes signal object actants, but (non-adverbial infinitives are negatived by peidio [â-]).
do- marks the theme for adverbial rhemes (exx. 65–7 Dim- is used to negate narrative converbs (see ex. 54).
below, rhemes italicized), and, by that token, for con- Irish gan is highly grammaticalized, non-inflecting (gan tú, gan é), and by that token not properly preposi- (61) Ni ddywedodd yntau ddim w e d i - e i - d d a r l l e n tional; it seems to include a ›non-existencxe‹ semantic … ›He didn’t say anything after reading it …‹ component. The precise details of the negation of the (62) Yna, a’i gyfaill Emrys yn ei ddanfon, i ffwrdd ag Irish converb are still obscure, but it does not seem that ef. ›Then, with his friend E. accompanying him on his gan- alone is the negativer in this case; may agus gan- be the negativing exponent (agus adverbializing con- (63) … an dá shuil sin d o - m o - c h r i a t h r ú … (76) Yr oedd yn lle digon hawdd loetran o’i gwmpas (64) Táim á ( = { a g - } + { a } ) d h é a n a m h seo anois h e b - i - n e b - e i c h - g w e l d . ›It was a place easy enough … ›I’m doing this now …‹ to loiter around without anyone seeing you.‹ (65) D’fhágadar ann é, agus is a n n dó ón am sin i (77) Ymddiheuraf am yr holl oedi yma h e b - a n f o n i- leith. ›They left him there, and there he is from that time ddiolch am y llyfrau. ›I appologize for that whole delay in (lit. not) sending to thank you for the books.‹ (66) Ní m a r s i n don ghasúr. ›Not thus the boy‹.
(78) … nó gur thit mé siar ar mo thóin … g a n cois (67) O í c h e dom agus mé a g - ó l sa teach ósta f á g t h a agam le cur fúm. ›… until I fell back on my seo … ›One night (lit. a night-me), I drinking at that backside … without a leg left to me to put under me.‹ (79) … toisc g a n Máire a - b h e i t h I mo theannta. (68) A r- f h i l l e a d h abhaile dom, … ›I returning ›… because of M. not being with me.‹ (80) … agus d’oibir sé seacht mbliana déag i Nua (69) Ta r é i s - t e a c h t ón Aifreann dúinn, … ›After Eabhrach g a n s c r í o b h ariamh go hÉirinn le hinse cá rabh a bheo nó a mharbh. ›… and he worked seventeen (70) I - m o - l u í anseo san fhéar dom, … ›Me-lying years in New York without ever writing to Ireland to say (71) … ag-dearcadh dom orthu, … ›… me-looking (72) Yr oedd ef a m - i d d y n t - b r i o d i y calan gaeaf hwn. ›He meant for them to marry (lit. was about for 17 The agens actant may be marked by o-: Wedi-priodi o’i phlentyn them to marry) on that All Saints’ Day.‹ oalf ›Her last child having married‹.
(81) Agus bhí sí thuas ansin a g u s - g a n - d u l - a i c i aon duine a fheicéail. (p. 153)18 ›And she was up therewithout going to see any person.‹ Exclamative nexus, nexus (or converb) as title, nexus asadjunct: (88) Ochón, a dúirt na daoine, fear bocht eile c a i l l t e ! ›Alas, said the people, another poor man lost!‹ (89) ›Mé p ó s t a ‹ ›ag seilg sna Rosa‹ ›»I married.« In Kate Robert’s fiction, ›typographical focussing‹ by segmental italicization (SHISHA-HALEVY 1995: §3.1.1) (90) Ní ar mo stól atáim anois ach i - m o - l u í an ar is consistently applied to the yn-/wedi- prefix – nucleus urlár … íse i n a - s u í le m’ais. ›It’s not on my stool that of the rhematic converbal syntagm, thus representative I am now but lying on the floor … she sitting at my – in a sense, the ›copula‹ – of the whole nexus, for nexus (82) ’Rydw i y n -dibynnu arnoch chi. ›I do rely on Dept. of Linguistics, The Hebrew University (83) Mi fasa’n haws gin Ann gredu ’mod i w e d i - of Jeru salem, Mt. Scopus, Jerusalem 91905, m a r w taswn i y n -sgwennu. ›It would have been easier for Ann to believe I have died if I had written.‹ (84) Rydych chi w e d i -marw. ›You have died.‹ Abstract. Dieser gelegentlich abschweifende Beitragmit einer gemischten metalinguistischen und metameta-linguistischen Perspektive diskutiert die deskriptive Angemessenheit und die Vorteile, den Begriff converb (adverbal, adnominal to non-specific nuclei) in der grammatischen Analyse des Irischen und Wali -sischen anzuwenden. Es handelt sich um einen 1995 (85) Yna, a’i gyfaill Emrys y n - e i - d d a n f o n , i ffwrdd wiederentdeckten und allgemein verwendeten Begriff, ag ef. ›Then, with his friend E. accompanying him on der in der allgemeinen Linguistik ein verborgenes his way, he left (lit. away with him).‹ Dasein führte, seit er 1902 für die altaische Sprachwis- (86) Agus a dhroim i o m p a i t h e liom, leag sé siar senschaft geprägt wurde und 1951 von H.J. Polotsky auf arís leath uachtarach a choirp. ›With his back turned to die äthiopischen Sprachen angewandt wurde. Das me, he laid back again the upper half of his body.‹ converb, so wird gezeigt, scheint nicht weniger grund- (87) Polláin a g u s - i a d - l á n le uisce. ›small pools legend für die linguistische Beschreibung zu sein als der Infinitiv (mit dem es unverkennbar Berührungspunkteaufweist).
Pentti Aalto: Studies in Altaic and Comparative Philo - John Haiman & Sandra Thomson (eds.): Clause Com- bining in Grammar and Discourse. Amsterdam 1988.
Gideon Goldenberg: ›Tautological Infinitive‹, Israel Eric P. Hamp: ›Inordinate Clauses in Celtic‹, in: Claudia Oriental Studies 1 (1971), 35–85.
Corum et al. (eds.), You Take the High Node and I’ll Take the Low Node: Papers from the Comparative Gideon Goldenberg: ›The Semitic languages of Ethio - Syntax Festival, Chicago 1973, 229–51.
pia and their classification‹, Bulletin of the School of Oriental and African Studies 40 (1977), 461–507.
Martin Haspelmath & Ekkehard König (eds.): Converbs in Cross-Linguistic Perspective: Structure and Mean- Gideon Goldenberg: ›Semitic Linguistics and the Gen- ing of Adverbial Verb-Forms – Adverbial Participles, eral Study of Language‹, Israel Oriental Studies 20 Gerunds. Berlin/New York 1995.
GRØNBECH 1979Kaare Grønbech: The Structure of the Turkic Languages 18 The two last exx. are from Malachy McKenna (ed.), Seanchas (translated by T.R. Krueger). Bloomington 1979.
Rann na Feirste, Baile Atha Cliath 2005, pp.120, 153 respec-tively. I am indebted for them to my student, Ms. Orit Eshel.
Martin Haspelmath: ›The Converb as a Cross-Linguis- Ariel Shisha-Halevy: ›Structural Sketches of Middle tically Valid Category‹, in: HASPELMATH & KÖNIG Welsh Syntax (I): The Converter Systems‹, SCelt 29 Martin Haspelmath (ed.): Coordinating Constructions. Ariel Shisha-Halevy: ›Literary Modern Welsh Narrative Grammar: Two Features Described‹, Journal of Celtic Linguistics 6 (1997), 63–102.
Nikolaus P. Himmelmann & Eva Schultze-Berndt (eds.): Secondary Predication and Adverbial Modifi- Ariel Shisha-Halevy: Structural Studies in Modern cation: the Typology of Depictives. Oxford 2005.
Welsh Syntax: Aspects of the Grammar of Kate Roberts. Münster 1998. [Studien und Texte zur Kel- Konstantin I. Kazenin & Jakov C. Testelets: ›Where Co- ordination Meets Subordination: Converb Construc- tion in Tsakhur‹, in: HASPELMATH 2004, 227–40.
Ariel Shisha-Halevy: ›Structural Sketches of Middle Welsh Syntax (II): Noun Predication Patterns‹, SCelt Ekkehard König: ›The Meaning of Converb Construc- tions‹, in: HASPELMATH & KÖNIG 1995, 57–95.
Ariel Shisha-Halevy: ›Celtic Syntax, Egyptian-Coptic Bernd Kortmann: ›Adverbial Participial Clauses in Syntax‹, in: Monika Hasitzka et al. (Hgg.), Das Alte English‹, in: HASPELMATH & KÖNIG 1995, 189–237.
Ägypten und seine Nachbarn: Festschrift Helmut Satzinger, Krems 2003, 245–302.
Geoffrey L. Lewis: Turkish Grammar. Oxford 1967.
Ariel Shisha-Halevy: ›Juncture Features in Literary Johannes Lohmann: Philosophie und Sprachwis- Modern Welsh: Cohesion and Delimitation – Prob- lematik, Typology of Exponents and Features‹, ZCP John Myhill & Junko Hibiya: ›The Discourse Function of Clause Chaining‹, in: HAIMAN & THOMSON 1988, Ariel Shisha-Halevy: ›Juncture Features in Shenoutean Coptic: Linkage and Delimitation‹, in: Mat Im- merzeel & Jaques Van Der Vliet (eds.), Coptic Stud- Vladimir P. Nedjalkov: ›Some Typological Parameters ies on the Threshold of a New Millennium: Proceed- of Converbs‹, in: HASPELMATH & KÖNIG 1995, 97– ings of the Seventh International Congress of Coptic Studies, Leuven/Paris/Dudley, Mass., 2004, 155–75.
Hans J. Polotsky: Notes on Gurage Grammar. Ariel Shisha-Halevy: Topics in Coptic Syntax: Struc- Jerusalem 1951. [Oriental notes and studies pub- tural Studies in the Bohairic Dialect. Leuven 2007.
lished by The Israel Oriental Society; 2] Ariel Shisha-Halevy: ›On Conversion, Clause Ordina- Hans J. Polotsky: ›Syntaxe amharique et syntaxe tion and Related Notions‹, in: Gideon Goldenberg & turque‹ in: Atti del Convegno Internazionale di Studi Ariel Shisha-Halevy (eds.), Proceedings, Ancient Etiopici, Roma 1960, 117–21.
Egyptian, Neo-Semitic – Methods in Linguistics: Workshop in Memory of H.J. Polotsky, Jerusalem G.R. Rahmati: Die Hilfsverben und Verbaladverbien im Altaischen. Berlin/Leipzig 1928.
Ariel Shisha-Halevy: ›The Converb in Egyptian and Gustav J. Ramstedt: Über die Konjugation des Coptic: a Note‹, Proceedings of NACAL 35 (San An- Khalkha-Mongolischen. Helsingfors 1902.

Source: http://ling.huji.ac.il/Staff/Ariel_Shisha-Halevy/docs/Shisha-Halevy,%20A.%20-%202010%20-%20Converbs%20in%20Welsh%20and%20Irish%20-%20A%20Note.pdf