4. Definitions and previous research on collaborative constructions
4.3 Previous research
4.3.1 Research on collaborative completion
Different researchers have used various terms for what here is called completion. Under the heading of “Collaborative built sentences”, Sacks (1995(I): 145) discusses that it is not
unheard of for two persons to collaborate to produce a single sentence. The normal way that is done, however, is that, say, one person produces an almost complete sentence and finds himself searching for a last word or a last phrase which he can’t find, and then the other offers it.
Sacks also focuses on an example from a musical where conversational behaviour of this type is used and he calls it “lovers’ talk”, when each one “produces part of a sentence, which the other then may complete” (ibid.). He concludes that it is
an obvious device to show, through this playing with the syntactic features of an utterance, that these people are close to each other. They’re a unit” (ibid.).
Sacks treats collaborative productions as social actions; the possibility of building a sentence together in conversation is at the same time a possibility for jointly constructing a social unit. When two or more people produce something that normally is done by one person, it shows
a kind of extraordinary tie between syntactic possibilities and phenomena like social organization. That is, an extremely strong way that these kids go about demonstrating that, for one, there is a group here, is their getting together to put this sentence together, collaboratively (ibid.).
The largest contribution and the most intensive work on collaborative production is produced by Lerner (1991, 1994, 1996, 2002, 2004a and 2004b). He defines completion which he calls either sentence-in-progress (1991, 1994, 1996) or pre-emptive completion (2004a) as follows:
In conversation, the pre-emptive completion of one speaker’s turn-constructional unit ... by a subsequent speaker can operate on that unit in a way that transforms it’s production into a sequence – a collaborative turn sequence. . .
Here a recipient responds to a prior speaker, not by waiting until completion to act, but by pre-empting that completion as a method of responding (Lerner 2004a: 225).
Most of Lerner’s work concentrates on what he calls compound turn constructional unit, formats where the first component is a strong indication of the second one (1991,1996). Normally, there would be a transition relevance place when the whole compound unit has been produced but in these collaboratively produced utterances the speaker change will sometimes occur when the first part has been uttered and the second speaker then anticipates what kind of syntactic construction is needed to complete the unit. Examples of this are sentences having “if” or “when” in the first part which will project “then” as a final component. Lerner (1994; 1996) has also explored completions which are strongly projectable, e.g. in an instance where the first speaker has started a list and the second speaker concludes it. Lerner’s focus on collaborative built sentences is solely syntactic, even though the formats he explores ,,are not all syntactic in nature, but range from pragmatic formats (lists) to semantic relation (contrasts) to interactional activities (doing disagreement)” (Szczepek 2000a: 4). He sees the sentence
production as “an interactional achievement” in itself and the co-production of a sentence as displaying aspects of “an interactionally relevant syntax” (Lerner 1991: 441).
In two separate articles on English, Szczepek (2000a and 2000b) explores both formal and functional aspects of collaborative non-competitive productions of syntactic constructions and prosodic units. Her definition of collaborative production is different from Lerner’s because she does not only claim syntactic continuation but also prosodic and pragmatic ones (Szczepek 2000a: 10). The second part of a collaborative production is therefore necessarily a continuation of the first component, syntactically, prosodically and pragmatically.
Helasvuo (2004) has especially studied completions involving the co-construction of one clause in Finnish conversations, and her main interest is to describe the syntax of completions in conversation. She divides the co-constructions into four categories according to the syntactic features which she analyses in more detail than the researchers mentioned above. Obviously, this approach gives a clear picture of the syntactic formats in collaborative constructions, but leaves out pragmatic and prosodic features which do not fall within the scope of her research.
In his studies on Swedish, Bockgård (2004) deals with three types of collaborative productions, i.e. completion, designedly incomplete utterances which require what Bockgård calls “a gap-fill-in-answer” (op.cit.: 282) and extensions (i.e. other-extensions). He focuses on the syntactic structure of these productions, especially on the internal syntax, with an emphasis on the ‘second move’, i.e. the second speaker’s act upon the preceding utterance. From Bockgård’s point of view, every second move has two indispensable factors; it has to be a syntactic continuation of A’s utterance, and it has to be connected pragmatically to the preceding utterance. This requirement goes for all the three phenomena Bockgård analyses in his thesis (103). Bockgård’s (2004) definition of completion rests upon whether B adds his utterance to an open or closed expression or, in other words, whether A has reached a potential TRP or not. Completion can, from Bockgård’s viewpoint, either be an induced completion, where there are problems in A’s utterance that result in B’s joining in, or a non-induced completion, where nothing seems to trigger B’s action (104).
Ferrara (1992) differentiates between three types of completions in English. These are predictable utterance completion, helpful utterance completion, and invited utterance
completions where the last one corresponds to Bockgård’s gap-fill-in-answers. Ferrara has been criticised for treating these categories as exclusive, when they are not. A completion can, for
instance, be an invited utterance and a helpful one at the same time (see Szczepek 2000a: 4); the first does not exclude the second. Ferrara focuses on syntax but leaves out prosodic aspects. Günthner’s (2012) focus is on constructions in German where the second speaker completes utterances in the making by adding dass-clauses (E. ‘that-clauses’). This collaborative
construction is possible because the partners in the conversation rely on cognitive routines on which the social action is based.
Howes (2012) combines a corpus analysis, experiments and theoretical modelling to explore what she calls compound contributions, i.e. how they are used and how they affect the conversation.11 Her evidence shows that compound contributions are frequent in different media, both in dialogues and in written texts, in fact „3–10% of all contributions in dialogue being continuations by one person of another‘s prior contribution and 10–24% being continuations of one‘s prior contribution“ op.cit.: 201).
The abovementioned research focuses on English, German, Finnish and Swedish. It is interesting to look at collaborative production in a language which does not only represent a difference in culture and interactional practices, but also gives an account of TCU constructions in a verb-final language. That is what Hayashi (1999) and Lerner and Tagaki (1999) have done. One of Hayashi’s (op.cit.: 495) findings is “that projection in Japanese is done more bit-by-bit than English” and that “syntax by itself is not as much of a resource for projection and therefore for co-participant completion” because the clause in Japanese seems to have a looser syntactic organisation than in e.g. English. Lerner and Takagi (op.cit.: 73), who also worked on Japanese data, concluded that an “[a]nticipatory completion of compound TCU structures found across languages furnish evidence of participants’ orientation to disparate syntactic structures of utterance production for similar features of turn-construction” (original emphasis overlooked).