Constraints on the Distribution of English Pronoun Case Forms in Coordinates

This abstract summarises findings from a detailed written survey of English pronoun case (cf. Quinn 2002). The results of the empirical study indicate that the distribution of English pronoun case forms in coordinates is difficult to account for in a purely case-based approach, but is nevertheless tightly constrained. Although the pronoun case patterns found in coordinates may vary considerably from speaker to speaker, the individual speaker patterns attested in the survey suggest that the distribution of pronoun case forms in coordinates is primarily influenced by the factors given in (1)-(3).

• (1) The GF of the coordinate
  • (a) Pronouns in SUBJECT coordinates are more likely to be nominative than pronouns in coordinates functioning as the OBJECT of a verb or preposition.
  • (b) Coordinated pronouns are more likely to be objective when the coordinate is the OBJ of a verb than if it is the OBJ of a preposition.
• (2) The position of the pronoun within the coordinate
  • (a) For 1sg, me is favoured in initial conjuncts, and I in final conjuncts.
  • (b) For non-1sg, he, she, we, they are favoured in initial conjuncts, and him, her, us, them are favoured in final conjuncts.
  • (c) they is more readily accepted in final conjuncts than me, he, she, we.
• (3) A general tendency towards the use of the objective pronoun forms me, him, her, us, them in coordinates. This tendency is stronger for non-1sg pronouns than for 1sg, and seems to be most advanced for 1pl.

I argue that the trends outlined in (1)-(3) are best modelled in a constraint-weighting approach where the cumulative weight of constraint violations determines the probability of occurrence of a particular variant (cf. Guy 1997, Mohanan 1998, Jäger and Rosenbach 2003). In order to account for the pronoun case patterns observed in the survey, we need to posit one or more case constraints that allow us to distinguish between subjects and objects, as well as between objects of verbs and objects of prepositions. I propose that these case constraints compete with two Relative Positional Coding constraints (4)-(5) and a set of Invariant Strong Form constraints (6).

• (4) Relative Positional Coding 1: A pronoun in initial conjunct position must be gracile and a pronoun in final conjunct position must be robust. Robust pronoun forms have rhymes of greater feature-geometric complexity than their gracile counterparts, i.e. they contain a diphthong or VC sequence rather than a lone monophthong (cf. Dogil and Luschützky 1990, Kenstowicz 1994, Bock 1982, and Quinn 2002 for further discussion of phonological complexity and its role in linear ordering preferences).

  The set of gracile pronoun forms comprises me, he, she, we, they.
  The set of robust pronoun forms comprises I, him, her, us, them.

• (5) Relative Positional Coding 2: A pronoun appearing in the final conjunct of a pronominal coordinate must be more robust than the pronoun in the initial conjunct.

  In the set of gracile pronoun forms, they is more robust than me, he, she, we, because its rhyme has a greater feature-geometric complexity.

• (6) Invariant Strong Form constraints: The morphological form of a strong pronoun must be invariant in all contexts (cf. Benveniste 1974, Kayne 1975, Cardinaletti and Starke 1995 and 1999, and Quinn 2002 for detailed discussions of the weak-strong pronoun distinction, and for evidence that coordinated pronouns in English are strong rather than weak).

  Since the trend towards an invariant form varies from pronoun to pronoun for most of the speakers surveyed, we need to assume a separate Invariant constraint for each pronoun.

All constraints are violable and have some weight for all speakers. The total number of violation points incurred by a candidate is determined by the combined weight of all the constraints it violates. Variation between speakers is captured by assigning different weightings to the different constraints. Variation within the speech of an individual occurs where the demands of different constraints clash, and two or more candidates incur a similar number of violation points.

* I would like to thank Peter Peterson for offering me the opportunity to participate in the workshop on coordination, and Joan Bresnan for drawing my attention to Gerhard Jäger and Anette Rosenbach's work on cumulativity effects in constraint evaluation.

References

• Benveniste, Émile. 1974. L'antonyme et le pronom en français moderne. Problèmes de linguistique générale II, 197-214. Paris: Gallimard.
• Cardinaletti, Anna and Michal Starke. 1995. The tripartition of pronouns and its acquisition: Principle B puzzles are ambiguity problems. NELS 25, Volume 2: Papers from the Workshop on Language Acquisition and Language Change, ed. by Jill N. Beckman, 1-12. Amherst, MA: GLSA, University of Massachusetts.
• Cardinaletti, Anna and Michal Starke. 1999. The typology of structural deficiency: a case study of the three classes of pronouns. Clitics in the languages of Europe, ed. by Henk van Riemsdijk, 145-233. (Empirical approaches to language typology 20, Eurotyp 5) Berlin: Mouton de Gruyter.
• Dogil, Grzegorz and Hans Christian Luschützky. 1990. Notes on sonority and segmental strength. Rivista di Linguistica 2. 3-54.
• Guy, Gregory R. 1997. Violable is variable: optimality theory and linguistic variation. Language Variation and Change 9. 333-347.
• Jäger, Gerhard and Rosenbach, Anette. 2003. The winner takes it all --- almost: cumulativity in grammatical variation. Ms, University of Potsdam and University of Düsseldorf.
• Kayne, Richard S. 1975. French syntax: the transformational cycle. (Current studies in linguistics series, 6) Cambridge, MA: MIT Press.
• Mohanan, K.P. 1998. Unpacking optimality theory: non-monotonicity as transformations, rule ordering and constraint ranking. Paper presented at the ALS Conference, 4 July 1998, Brisbane.
• Quinn, Heidi. 2002. The distribution of pronoun case forms in English. Unpublished PhD thesis, University of Canterbury, New Zealand.