Sentence processing in an artificial language: Learning and using combinatorial constraints

Abstract

A study combining artificial grammar and sentence comprehension methods investigated the learning and online use of probabilistic, nonadjacent combinatorial constraints. Participants learned a small artificial language describing cartoon monsters acting on objects. Self-paced reading of sentences in the artificial language revealed comprehenders' sensitivity to nonadjacent combinatorial constraints, without explicit awareness of the probabilities embedded in the language. These results show that even newly-learned constraints have an identifiable effect on online sentence processing. The rapidity of learning in this paradigm relative to others has implications for theories of implicit learning and its role in language acquisition.

Figure 1

Sample picture and accompanying sentence, with English gloss and translation.

Figure 2

Frequency in each training block of sentences containing conjunctions of the three critical elements. Monsters (yeen, pim, gled) are shown at left, objects (sarp, clate, gorm) across top of grid, and verbs (dak[torch, i.e., breathe fire on], grah[lift], veek[whip]) are nested within each monster in the grid. For example in each training set 7 sentences contained the frequent veek+pim+sarp conjunction shown in Figure 1.

Figure 3

Participants' mean self-paced reading times by sentence type. Word6 was read faster when the sentence contained a frequent conjunction. Bars show standard errors.