In The Discourse Basis of Ergativity published in Language in 1987, John W. Du Bois proposed a theory which stated that (p. 850)

[universally] the distribution of new information vs. old information [as well as the distribution of lexical vs. non-lexical arguments] is correllated with the ergative/ absolutive structural opposition.

According to him,

language-internal phenomena as basic as the system of grammatical relations can be structured by forces arising out of discourse.

Basically, the S=P alignment found in ergative languages is said to be a manifestation of the pattern in discourse, which also shows S=P alignment of new information (pragmatic dimension of the Preferred Argument Structure) and noun phrases (lexical dimension). Note that Du Bois gives several examples for languages (next to the one he observed) that also seem to show that alignment, among others German, Quechua and English (p. 839).

This is where I come in: In a recent homework, I tried to find supporting/contradictory evidence in another corpus. My own data did in the end confirm Du Bois' theory.

However, a former professor now "published" a manuscript, in which he argues that (Haig/Schnell p. 40)

outside Du Bois' original data, evidence in support of an S=P unity in discourse are strikingly absent

and that

the claim that cross-linguistically, connected discourse exhibits a characteristically ergative profile, such that intransitive subjects and objects pattern alike, is in fact an artifact of a highly marked corpus.

For this reason, I would like to ask you if you yourself have reason to believe that either of the two positions is correct. Ideally, try to confirm/refute Du Bois' theory on a Preferred Argument Structure based on findings of your own.

I can see how this question might provoke a discussion rather than definite answers. But as a clear, scholarly response is what I am aiming at here, feel free to give advice on how I could improve this question!

Both articles are available online.

Du Bois (1987): http://www.jstor.org/stable/415719

Haig/Schnell (2013): http://www.academia.edu/3403768/The_discourse_basis_of_ergativity_revisited (Note that this paper is still marked "unpublished manuscript"!)

The sample data I worked with was taken from the corpus developped in the project described here: http://www.hrelp.org/grants/projects/index.php?projid=131

  • 1
    S=P alignment is not a transparent term. What is S? What is P? What is =? What is aligned, in which dimension, in what structure? If it's sposta describe ergative usage in real ergative languages, it really needs to be more than a totally abstract formula.
    – jlawler
    Aug 25 '13 at 18:54
  • S=P alignment means that S and P are marked the same. Of course, alignment patterns can (and often do) vary across different constructions. Linguists are unfortunately not always explicit about the constructions they're describing, although what they mean is most often either case marking or agreement / head marking.
    – Fryie
    Aug 25 '13 at 22:47
  • S and P are standard terms in the alignment literature, although Haspelmath (2011) shows that there are at least 3 contrasting definitions of the terms. Note that many authors (most prominently, R.M.W. Dixon) use O instead of P (which is used e.g. by Comrie), but this is merely a notational variant, as well as Lazard's X, Y and Z. In rough terms, S is the single intransitive argument, while A and P/O are the more agentive and the more patientive arguments of the transitive clause, respectively.
    – Fryie
    Aug 25 '13 at 22:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.