Potsdam 22-25 September 2004
Response to Peter Siemund’s paper “Independent Developments in the Genesis of Irish English”. Una Cunningham, Högskolan Dalarna, Sweden.
In the introduction, this paper explains that in the study of IrE it is usual to look for explanation of non-standard features in the superstratum, that is the retention of features of earlier dialects of English, and the substratum, that is the transfer of features from Irish. Both of these efforts involve the comparative method of historical linguistics to establish the historical source of morphosyntactic forms. P.S. suggests that there might be effects of the contact situation in itself that can have given rise to at least some of the morphosyntactic peculiarities of IrE. These would be things like
imperfect learning, and this is something that P.S. returns to later, overgeneralisation, speaker creativity, and significantly, pressure from linguistic universals. In fact he goes as far as to say that an unstable linguistic situation such as that found in the shift from Irish to English (1700-1900) will
“inevitably trigger the activation of linguistic universals”.
Section 2 of the paper delves into the advantages and limitations of the retention/transfer debate.
The idea in this approach is, according to Peter Siemund, that every morphosyntactic property of a contact variety like IrE has a traceable history as either originating in earlier English dialects or Irish. Surface forms are forms rather than underlying structures are compared. The example of the after perfect exemplifies a construction with an Irish source. In this case the transfer is of the function of a morpheme which is projected onto an English morpheme. <this seems to me the case in most kinds of transfer, cf can I be with>.
P.S. claims not to have been able to find a single direct loan from Irish into IrE in the literature or in his own data. Filppula et al in this colloquium have, quoting from Thomason and Kaufman, claimed that this is a fallacy and that there is an underreporting of loans from Irish to IrE. P.S. also claims that the borrowing of morphemes, particularly of grammatical morphemes is an important
diagnostic of this kind of rapid, contact induced language shift. This kind of borrowing is also found in the speech of infant bilinguals when they clothe the grammatical morphemes of their dominant language in the lexical material of their weaker language. (först climber man upp sen kan man slida ner) But this phase is rapidly abandoned in the speech of these infant bilinguals and there may not be a need for it in a shift situation that has taken many generations. One might surmise that pressure from a strict corps of prescriptive schoolmasters might have caused this to disappear.
In ex 3 and 4 subordinating and is examined. P.S. describes this as the transfer of a function from an Irish morpheme to an English morpheme, including the bundle of morphosyntactic properties. The gloss of the Irish sentence 4 suggests a totally different analysis. This makes me think of the question “what are you at?” which would be answered with a progressive – “what was he at when /that he fell?” Surely this is more than the transfer of a morpheme’s function in the and/agus?
Rather the transfer of the structure itself?
In the case of the medial object perfect, (p4) P.S. claims that the alleged transfer from Irish to IrE must have been subtle because of the differences in construction used to encode the meaning. He argues that the distance is great since a locative construction is used in Irish and a verb in English to encode possession. But is this not also the case in other constructions? Is it not the use of a
possessive construction to express the perfect that is transferred, i.e. a semantic transfer rather than a morphosyntactic one? P.S. argues that the case for influence from Irish is weakened by the number of cognitive steps required to make the IrE construction from the Eng one. But surely the point is not to go from Irish to IrE in a certain number of steps, but to show why IrE has the kinds of specific constructions it has.
Looking at the case of the omission of pronominal arguments in subject and object position in IrE, ex. 7, 8, 9, P.S. finds it a challenge to the traditional approach to have structural properties being transferred without there being any carrier morphemes involved. This is perhaps parallel to the case of the Swedish English children who ask “can I be with” or conversely the Swedish song with the somewhat sinister lyrics “Can I follow you home?”. Why should transfer be restricted to
morphemes. There are other ways for languages to influence each other.
P.S. points out the limitation of the traditional analysis that any non-standard features of IrE that cannot be traced to either Irish or earlier dialects of English are ignored. He mentions constructions like double perfects, infinitives instead of –ing forms and the use of nominative pronouns as
subjects of non-finite clauses. This latter case is given quite a bit of attention in this paper. The traditional method, writes P.S. has very little to say about pronouns like in ex 10. Might this not be a case of hypercorrection? In the face of uncertainty about the use of the subject and object
pronouns, prescriptivism generally comes down on the side of the subject, so go for it…? Other examples mentioned are e.g. marking of possession which P.S. claims has not led to noticeable influence from Irish on English. But what about e.g. “he has a fierce temper on him” with a kind of
2
double marking? Is not this just the kind of hybrid form P.S. finds lacking? Just in the situation of extensive bilingualism and imperfect learning?
In the third section, P.S. introduces the universalist approach. The point of this is to identify the properties of IrE that are cross-linguistically relevant and not idiosyncratic. P.S. argues that more than just transfer and retention played a role. He mentions code switching, and first and second language acquisition strategies in this respect, such that the contact situation itself left linguistic traces in IrE. P.S. distances himself from analysing IrE as a pidgin or Creole situation, on the grounds that the social and linguistic conditions for the emergence of such are not met. Other writers (e.g. Corrigan 1993) have no such qualms. While IrE cannot, as P.S. points out, be said to have complete grammatical subsystems of Irish, there is certainly enough there to make the average 11 year-old learner of Irish at school feel that the language is not foreign, but that the structures and expressions of IrE are clothed in Irish words.
P.S. talks about applying both the concept of UG and of typological universals to the study of IrE.
He talks of the deficient primary data of non-native models being compensated by ug. Well young children who are in the process of parameter setting will not be set on the way to English-like word order if their input suggests a setting of OV. Yet as P.S. points out there is no evidence of Irish word order in IrE. But if we are to accept the role of UG in L1 acquisition, we need to consider that word order errors are never found in the language of monolingual infants. L2 learners do however make such errors (their parameters are already set).
In the fourth section P.S. introduces his data, and in the fifth section he looks at two “case studies”:
medial object perfects and nominative pronouns. In the first case, P.S. argues that it is not plausible to assume that this construction is a transfer from Irish into English because it is well attested in earlier dialects of English. Further he claims the Irish construction is a passive-cum-possessive construction where possessivity is expressed by a locative phrase which he suggests makes transfer complicated. Compare this construction with the German for I have written a letter. If that is what the basis for the earlier dialects is it is far removed from the possessive feeling of the IrE
construction. Compare this with e.g. “you have him half reared” said of a big baby. Ex 15 is parallel with I have it here, or where do you have it?
3
In the case of have got in ex 19, there is a different meaning. This is the perfect of to get something done. It means “I’ve managed to write it.” It is even conceivable, according to my intuitions that we can have an IrE construction like It was a lot of work but I’ve it (got) done now.
As for the examples in 20 and 21, how would the following fit in:
He had the children in bed by seven He had the children put to bed by seven
This use of a temporal construction makes these sentences feel more convincing. How does this fit in with Peter Siemund’s grammaticalization process? I think it supports it.
On then to the nominal subject. Peter Siemund’s discussion about prescriptive norms is key here. In many situations of confusion, the nominal pronoun seems to be the recommended, and thus the safe option. I think there could be an element of hypercorrection here. Incidentally, how do ex 3 and 4 fit in here? (he fell and him crossing the bridge).
On p 17, P.S. accounts for this construction in terms of markedness – the nominative pronoun would thus be preferably to the accusative on the grounds that the latter, being less common is marked. It is not rather that the accusative pronoun is less common simply because sentences with expressed objects are less common than sentences with expressed subjects?
Finally, I would wholeheartedly agree with P.S. that the results of first and second language acquisition studies can be most illuminating for our understanding of the processes behind the development of IrE.
Refs
Corrigan, Karen 1993. ‘Hiberno-English syntax: Nature versus nurture in a creole context’, Newcastle and Durham Working Papers in Linguistics 1, 95-131.)
4
