The main focus of this section is on Middle Welsh examples of the use of gwneuthur ‘do’ as an auxiliary in a periphrastic construction. To counteract any suggestion that the construction in Welsh was influenced by English and to show that it is instead a feature of Brythonic, I also outline the parallel constructions in Cornish and Breton. Welsh gwneuthur, Cornish
gruthyl and Breton ober, all meaning ‘do, make’ as lexical verbs, also function as auxiliaries
in conjunction with a verbal noun. Gwneuthur, gruthyl and the inflected forms of ober are cognate, and have the root gr- or gn-, while the form ober itself, as a verbal noun, is suppletive from Latin opera ‘work’ (Elmar Ternes, personal communication 6/2/2015).
19 The word order is considered ‘abnormal’ from the point of view of Modern Welsh (D. S. Evans 1964: 179-
102 Although the use of the periphrastic construction with auxiliary ‘do’ can be assumed to have changed in some ways during the period from early Brythonic Celtic to the modern forms of the languages, it is similar in structure in all the Brythonic languages for which we have evidence, and it can be regarded as one and the same periphrastic construction continuing from the Common Brythonic period. However, the middle periods of these three languages is the earliest for which we can find suitable evidence of periphrasis, since the oldest surviving examples of these three languages are too short for syntactic analysis. Since some proponents of the Celtic Hypothesis, for instance Preusler (1938), McWhorter (2009) and Filppula, Klemola and Paulasto (2008: 49-59), use examples from modern languages to support their arguments, I begin my discussion with the modern Brythonic languages, Welsh and Breton. Cornish, of course, died out and so ‘modern times’ for Cornish ends for practical purposes in the seventeenth century, although it held on longer in isolated pockets. Although Cornish is being revived, it is not an appropriate stage of the language for my purposes.
Modern Welsh
One reason for giving examples from Modern Welsh is that, as a spoken language, it might be considered to shed light on an earlier spoken form of the Brythonic language that is claimed to have influenced early English. However, not only would the intervening
millennium and a half have distorted this light immeasurably, but the use of Modern Welsh examples sometimes obscures the differences from the periphrastic construction in Middle Welsh. For instance, periphrasis with gwneud (the informal form of gwneuthur) in
conjunction with a verbal noun is common in only some registers of Modern Welsh. It is used freely in colloquial, spoken Welsh; where it is found in writing it is marked as being a
representation of the informal or spoken language. However, such periphrasis is not usually found in formal or written forms of contemporary Welsh. Proponents of the Celtic
Hypothesis who use examples from Modern Welsh to illustrate the Brythonic periphrastic construction need to be aware of the differences in its use in Modern Welsh from that of the Middle Welsh construction with gwneuthur. Nonetheless, providing Modern Welsh examples can be advantageous, considering the fact that some linguists may be more familiar with Modern Welsh than they are with Middle Welsh.
Looking at some of the modern grammars for language learners shows that periphrasis with
gwneud is a complex topic. For instance, King (1993), writing for language learners, says that
there are three ways of expressing ‘he bought/he did buy’ in Modern Welsh: brynodd e, naeth
e brynu and ddaru o brynu (1993: 136-7), which I gloss here:
(22) brynodd e buy-3SG.PRET he (23) naeth e brynu
103 (24) ddaru o brynu
happen-3SG.PRET he buy-VN
King explains that his list of translations is simplified (1993: 137), as is to be expected in a grammar book aimed at learners of spoken Welsh. While it might seem strange to begin my discussion with a simplified grammar for beginners, many other grammars ignore
periphrastic gwneud altogether. Awbery (1976: 13-9, 96-9), for instance, whose focus is on the periphrastic use of cael ‘get’ to convey the passive in Modern Welsh, also discusses periphrastic constructions with bod ‘be’ and a few other verbs, but not those with gwneud. Stephen Williams (1980: 72-6, 94-5,117-20), too, discusses periphrasis with bod, but not with
gwneud. The reason for the absence of reference to the periphrastic construction with gwneud
in these two discussions of Modern Welsh is that they describe the grammar of standard literary Welsh (Awbery 1976: 2, S. J. Williams 1980: v), which takes the Welsh translation of the Bible of 1588 as its standard (S. J. Williams 1980: 85). Although examples of the
periphrastic construction with gwneud are common in spoken Modern Welsh and can be found in informal forms of the contemporary written language, they are usually avoided in formal, literary Welsh. Periphrasis with gwneud can, however, also be used in formal Welsh if the lexical verb is emphasized (Elwyn Hughes, personal communication, 20/1/2015), a point that both Awbery and Williams omit to mention. Since fronting can be used in Welsh to emphasize most sentence constituents, but the unmarked position for verbs is clause-initial, to be able to emphasize the verb the verbal noun takes the clause-initial position of the inflected verb, leaving the inflected auxiliary gwneud to follow later in the clause (Borsley, Tallerman & Willis 2007: 126-8). This emphatic use of gwneud is paralleled in English by the use of do- periphrasis for emphasis, although the word order is naturally reversed. Like English do,
gwneud can also be used as a pro-verb, as the ‘code’ in ellipsis (Rouveret 2012). Unlike do-
periphrasis in English, therefore, periphrasis with gwneud in formal Welsh occurs in restricted contexts.
It is, therefore, important to clarify that some forms of periphrasis are not available in all language registers of Modern Welsh and that there may also be regional variations in which particular verbs are used as auxiliaries in such periphrasis. ‘Do’-periphrasis in Modern Welsh is, therefore, not only a common construction, but it is also a complicated one in terms of when it is used and how it is realised. Proponents of the Celtic Hypothesis that rely on examples from Modern Welsh to draw a parallel with the construction in English would do well to acknowledge this, since they tend to give a ‘normalised’ version of the Welsh language.
Periphrasis with ober in Modern Breton
Press (1986: 128) gives three ‘conjugations’ in Modern Breton, one of which is periphrastic using ober ‘do’, for which he gives the following example with the Imperfect of ober to express the habitual in the past:
104 (25) dont a raent do Lannuon
come-VN PT do-3PL.IMPF to Lannion ‘They used to come to Lannion’ (My gloss based on his; his translation)
Borsley, Rivero and Stephens (1996: 54, 66) discuss periphrasis in Modern Breton in more depth and they gloss the following example of an inflected form of ober used as an auxiliary with a verbal noun:
(26) Lenn a ra Anna al levr read PT do.3SG.PRES Anna the book ‘Anna reads the book.’
It is clear that ‘do’-periphrasis is an established syntactic feature in the modern Brythonic languages. However, it has to be established what form the periphrasis in Brythonic took at the time its speakers were first learning to speak English and whether this form bore any resemblance to the early English do-periphrasis. If it did not, the case for Brythonic influence may be weakened. The fact that Modern English do-periphrasis has a similar structure to periphrasis with gwneud in informal Modern Welsh and ober in Modern Breton is indeed suggestive, but not sufficient to rest a hypothesis on without a demonstration that the modern languages are representative of the Brythonic language(s) spoken during the fifth and sixth centuries. After a brief outline of the construction in Middle Breton and Middle Cornish, I discuss the situation in Middle Welsh at more length.
Periphrasis with ober in Middle Breton
The Middle Breton periphrastic construction with ober is similar to the constructions in Modern Breton, Middle Cornish and Middle and Modern Welsh. As Willis (2007: 287-96) explains, the middle stages of these Brythonic languages are verb-second showing the abnormal word order, unlike Modern Welsh. This means that the periphrastic construction in Middle Breton has the word order verbal noun + particle + ‘do’ + rest of sentence. Although Willis does not specifically address the periphrastic construction with ober, he gives an example of it from Middle Breton:
(27) fezaff agra en holl tut sauant beat-VN PT+do.3SG the all people wise ‘she beats all the wise people’
(My gloss based on his; his translation)
The construction is used frequently according to Hemon (1975: §158), who gives the following example:
105 (28) mont de Velem a rezant-y
go-VN to Bethlehem PT do.3PL-they ‘they went to Bethlehem’
(My gloss; his translation)
Periphrasis with gruthyl in Middle Cornish
Periphrasis with gruthyl ‘do’ and a verbal noun is also a common syntactic structure in Cornish. In Middle Cornish, gruthyl ‘do’ can be used as a lexical verb meaning ‘do, make’ in an inflected form, as a verbal noun or even as a past participle. It can also be used as an auxiliary in conjunction with a lexical verb in the form of a verbal noun (Zimmer’s trans. of H. Lewis 1990: 46-7). Inflected forms of gruthyl as a lexical verb (29 and 30) and as an auxiliary (31) can be seen in the following examples from one of the fifteenth-century mystery plays edited and translated by Norris (1859: 6-7, 108-9):
(29) ny a ’d wra ty then a bry we PT you.OBJ do-3SG.PRES you man of clay
‘we make thee, man, of clay’ (Ordinale de origine mundi 69)
(30) ty ru ’m gruk pur havel thy ’s you PT me do-3SG.PRET very like to you
‘thou hast done me, very like to thee’ (Ordinale de origine mundi 88) (31) guelen a pren a wraf synsy
rod of wood PT do-1SG.PRES hold-VN
‘a rod of wood I do hold’ (Ordinale de origine mundi 1444) (My glosses; Norris’s translations)
A brief glance at Ordinale de origine mundi gives the impression that the periphrastic use of
gruthyl as an auxiliary occurs much less frequently than the use of inflected lexical verbs. On
the one hand, this text may well represent the spoken language more accurately than some other writings, since it is written in direct speech as drama. On the other hand, the reduction in the use of periphrasis could be because it, as is usual in drama, is written predominantly in the present tense. So, if the periphrastic use of gruthyl is more common in the preterite than in other tenses in Middle Cornish, as is also the case in Middle Welsh, the mystery plays may be unrepresentative of Middle Cornish writing in having a reduced frequency of the preterite tense. However, this text may be unrepresentative as evidence of the use of periphrasis with
gruthyl for another reason, since, according to Thomas (1992a: 346), the Cornish miracle
106 syntax of the English originals has influenced that of the Cornish texts. What date these putative English originals may have been and whether they showed evidence of do- periphrasis is too hypothetical to contemplate.
Fife contrasts the periphrastic use of gruthyl with the periphrastic use of gwneuthur in Welsh: ‘The Cornish structure is used to form the regular present tense of the VN, a use not possible with the Welsh DO’ (1990: 216). Although the ‘simple present’ of gwneuthur can be used as an auxiliary in Modern Welsh, it conveys the future tense; the present tense in Modern Welsh is conveyed by the present tense of bod ‘be’ with a preverbal particle and a verbal noun.
Periphrasis with gwneuthur in Middle Welsh
The use of periphrasis with gwneuthur in Middle Welsh is in some ways similar to that of Modern Welsh. On the one hand, in Middle Welsh and in informal Modern Welsh there are two ways of forming some tenses: an inflected lexical verb (the ‘short form’) or an inflected form of gwneuthur with a verbal noun for the lexical verb, and the choice between them appears to be optional. On the other hand, in Middle Welsh both the short form and the periphrastic construction sometimes occur in the same formal, literary text, something that does not occur in formal Modern Welsh as I explained above. Furthermore, there are some Middle Welsh texts that show no evidence of the periphrastic construction at all. Arguments in support of the Celtic Hypothesis that rely on Modern Welsh evidence alone are, therefore, potentially misleading.
Strachan (1909: 78-9) states that periphrasis with gwneuthur and a verbal noun is common in Middle Welsh, and he gives these two examples from Ystorya Gereint Uab Erbin, one of the romances associated with the Mabingion:
(32) a cherdet recdi yr coet a oruc y uorwyn and walk-VN before.her the wood PT do-3SG.PRET2 the girl ‘and the girl went to the wood’
(33) rodi penn y karw a waethpwyt y Enid give-VN head the stag PT do-3SG.PRET1.IMPERS to Enid ‘the stag’s head was given to Enid’
(My glosses; Strachan’s translations)
Examples (32) and (33) show the basic structure of the periphrastic construction in Middle Welsh of a verbal noun (VN), a verbal particle (PT) and an inflected form of gwneuthur. The construction also occurs in variant forms. For instance, it also occurs with more than one verbal noun governed by an inflected form of gwneuthur:
107 (34) dynessau a oruc Osla Gyllelluawr […] a dygrynnyaw yndaw
approach-VN PT do-3SG.PRET2 Osla Big-knife […] and surround-VN to-him ‘Osla Big-knife approached and surrounded him’
(Culhwch ac Olwen 1180–2) (My gloss; Strachan’s translation)
Another Middle Welsh sentence structure appears similar, although it is technically not periphrastic: Strachan notes that the verbal noun can be used without an inflected form of
gwneuthur as ‘an historical infinitive’ (1909: 79), as at (35). This omission of the auxiliary is
not simply scribal error. Fowkes (1991: 21-4) gives 28 examples from the Mabinogi of the verbal noun as the only verb in a clause and I have found additional examples of the verbal noun alone in Culhwch ac Olwen, for instance at lines 510–1, 920–1 and 1174, not counting examples where it is uncertain whether the verb form is the verbal noun or the uninflected verb stem for the 3rd person singular.
(35) dyuot y porthawr ac agori y porth come-VN the porter and open-VN the gate ‘the porter came and opened the gate’
(Culhwch ac Olwen 786) (My gloss; Strachan’s translation)
From my own analysis of Culhwch ac Olwen, I have found that a common pattern of usage in Middle Welsh prose narrative – and this is unlike the situation in formal Modern Welsh – is for a sentence with periphrasis to occur alongside one using the short form, at the overall rate of about one sentence with periphrasis to three or four sentences using the short form. For instance, example (36) has a sentence without periphrasis that is immediately followed by one with periphrasis (37):
(36) o hynny allan y gelwit Goreu mab Custenhin from that out PT call-3SG.PRET.IMPERS Goreu son Custenhin ‘from there on out he was called Goreu son of Custenhin’
(Culhwch ac Olwen 810–1) (37) guascaru a orugant wy y eu llettyeu
scatter-VNPT do-3PL.PRET2 they to their beds
‘they scattered to their beds’ (Culhwch ac Olwen 811)
This alternation is particularly striking when a lexical verb is used first as a verbal noun with an inflected form of gwneuthur (38) and then as the inflected verb itself in the very next sentence (39):
(38) dywedut a wnaethant y Arthur y ual y daruu udunt say-VN PT do-3PL.PRET1 to Arthur PT how PT happen-3SG.PRET to-them ‘they said to Arthur how it had turned out for them’
108 (Culhwch ac Olwen 825)
(39) Arthur a dywawt Arthur PT say-3SG.PRET
‘Arthur said’ (Culhwch ac Olwen 826)
These examples show that periphrasis with gwneuthur and a verbal noun exists in Middle Welsh. However, there are features of it that may be obscured when proponents of the Celtic Hypothesis rely predominantly on examples from Modern Welsh. Clearly, the context in which the periphrasis can occur differs in an important respect: the construction occurs optionally in formal literary texts in Middle Welsh alongside the short form, but not in Modern Welsh. The word order of the construction in Middle Welsh also differs and I shall have more to say about that in the next section. First, however, I wish to consider how representative the examples above are of the use of the periphrastic construction of
gwneuthur and verbal noun in Middle Welsh.
Taking Culhwch ac Olwen as a test case, my analysis shows that both the simple form and the periphrastic construction occur throughout the four sections I analysed (lines 1–200, 374– 600, 751–952 and 1057–1246 of the edition by Bromwich and Evans (1988)), although the relative frequency of the two constructions, short form or periphrastic, is not uniform across the four sections. A comparison of the numbers of periphrastic and non-periphrastic verbs shows this variation in frequency with the non-periphrastic form being favoured throughout. In the first section, there are 28 instances of the periphrastic construction with gwneuthur to 215 non-periphrastic. In the second section, there are 28 instances of periphrasis to 337 non- periphrastic. In the third, there are 35 periphrastic to 294 non-periphrastic. In the fourth section of this text, the ratio is 32 periphrastic to 211 non-periphrastic verb forms. In the 819 lines of Culhwch ac Olwen I analyse, the periphrastic construction occurs at about the rate of 15 examples to every 100 lines, while the non- periphrastic construction occurs at about the rate of 129 to every 100 lines. Table 4.1 below summarises these figures and those for the other texts I have analysed.
109
Table 4.1: The number of periphrastic and non-periphrastic verbs in a selection of Middle Welsh texts Notes:
1. In counting the number of non-periphrastic verbs, I have not included verbal nouns except where they take the place of a finite verb.
2. In counting the number of periphrastic constructions, I have not distinguished between the ‘abnormal’ and ‘mixed’ word orders.
As can be seen from the table, the periphrastic construction with gwneuthur occurs in Pwyll
Pendeuic Dyuet, the first of the Mabinogion tales, slightly less frequently than it does in Culhwch ac Olwen. However, how the periphrasis is used in both texts is very similar, and its
average frequency per hundred lines in Culhwch ac Olwen would be reduced if I included passages such as lines 201-373, which contains few verbs and possibly no periphrasis. The periphrastic construction with gwneuthur also occurs in the prose chronicle Brut y
Tywysogion. In the following two examples taken from the Red Book of Hergest edition of
Thomas Jones (1955), the use of gwneuthur as a lexical verb in the entry for the year 814 can be contrasted with the periphrastic use of gwneuthur with the verbal noun anuon ‘send’ in the entry for the year 1100–110220. (The construction in (40) is not affected by the variant
reading y gwnaeth.)
(40) Ac yna y bu daran vawr ac a wnaeth [y gwnaeth RT] llawer o loscuaeu and then PT was thunder great and PT do-3SG.PRET1 lots of fires ‘And then there was great thunder and it caused many fires’
(Brut y Tywysogyon, 814 CE)
20 Entries in the chronicle are often given two dates: the first is the date given in the text and the second is the
date determined from external sources. In this instance, the text gives the year for the entry as 1100, but the actual year was 1102.
Text Number of lines Number of periphrastic gwneuthur constructions