Language users as creatures of habit: A corpus-based analysis of persistence in spoken English

A probabilistic corpus-based model of syntactic parallelism

“…There are a number of possible explanations for this difference. Priming has been argued to be subject to distance-based decay (e.g., Gries, 2005;Szmrecsanyi, 2005;Reitter et al, 2006). This may be a relevant factor as prime and target are relatively close together in coordination (only separated by one word), while the mean distance between prime and target is larger for priming in non-coordinate structures within the same sentence, and even larger for priming between sentences.…”

“…First, the existing literature on priming in corpora (e.g., Gries, 2005;Szmrecsanyi, 2005;Reitter, Moore, & Keller, 2006;Jaeger, 2006aJaeger, , 2006bJaeger & Snider, 2007) reports that a variety of factors can influence priming, including the distance between prime and target, the type and genre of the corpus, and whether prime and target are uttered by the same speaker. Previous work has used multiple regression methods to study priming in corpora; this approach is particularly useful when several factors are confounded, as regression makes it possible to quantify the relative contribution of each factor.…”

“…Our studies build on previous corpus-based work which demonstrated parallelism in coordination (Levy, 2002;Cooper & Hale, 2005), as well as between-sentence priming effects. Gries (2005), Szmrecsanyi (2005), Jaeger (2006aJaeger ( , 2006b, and Jaeger and Snider (2007) investigated priming in corpora for cases of structural choice (e.g., between a dative object and a PP object or between active and passive constructions). These results have been generalized by Reitter et al (2006), who showed that arbitrary rules in a corpus can be subject to priming.…”

“…First, we found in Section 2 that priming effects were smaller as the size of the priming region increased (from coordinate structures to arbitrary structures within sentences, to arbitrary structures betw een sentences). Second, a number of authors (e.g., Gries, 2005;Szmrecsanyi, 2005;Reitter et al, 2006) found in corpus studies that the strength of priming decays over time (but not Jaeger, 2006b andSnider, 2007, who controlled for speaker differences). Intuitively, these two effects are related: by selecting a larger priming region, we effectively increase the time between the onset of the prime and onset of the target.…”

Gradient grammar: An effect of animacy on the syntax of give in New Zealand and American English

“…This idealization has been fruitful, but it ultimately underestimates human language capacities. Language users reliably and systematically make probabilistic syntactic choices from multidimensional information (Arnold et al 2000, Bresnan 2006, Gries 2003, 2005a,b, Hay and Bresnan 2006, Hinrichs and Szmrecsányi to appear, Jaeger 2006Roland et al 2005, O'Connor et al 2005, Rosenbach 2002, this volume, Strunk 2005, Szmrecsányi 2005. In a recent study of English speakers' syntactic choices with give-type verbs during spontaneous conversations, Bresnan et al (2007) present a multivariable, multilevel logistic regression model that can accurately predict the choices on unseen data.…”