Research demonstrates that when participants read a sentence about an agent in a certain location and then are showna pictured object, verification time is shorter whenever the pictured object matches the final object state implied by thesentence. Using a sentence-picture verification paradigm, we set out to investigate if the same pattern of results holdstrue when proprioceptive and kinesthetic experiences are considered. In three experiments participants read sentencesthat implied object state-changes as a function of the impact caused by differently weighted items (You drop a bowlingball/balloon on a tomato) followed by a pictured object in either a canonical (e.g., a round tomato) or a non-canonical (e.g.,a squashed tomato) state. The results showed that depictions of non-canonical objects showed the effect, but depictionsof canonical objects did not. Thus, representations of object states compete when non-visual features of the situation areimplied by the sentential context.

Many studies show that language comprehenders track multiple object state-changes during event comprehension. In this research, we investigate if comprehenders are sensitive to object state-changes of semantically similar objects. In Experiments 1 to 3, participants were asked to verify whether the object depicted in a picture was mentioned in the previously read sentence. Crucially, the picture either showed the original/modified state of an object that was mentioned in the sentence (Experiment 1) or not (Experiments 2 and 3). In Experiment 1, participants were faster to verify that an object was mentioned in the sentence if its pictured state matched the state implied by the sentence. In Experiments 2 and 3, participants were slower to verify pictured objects (e.g., cake) after reading a sentence that implied multiple states for a semantically similar (e.g., egg) rather than dissimilar (e.g., thermometer) object. Thus, comprehenders are sensitive to object state-changes of semantically similar objects.

A growing evidence from a sentence-picture verification paradigm suggests rapid integration of implied perceptual context during sentence processing. This evidence, however, is usually criticized for not providing a strong test of the mechanisms underlying such integration. We addressed this question in relation to the situation describing different sources of light. We hypothesized that if comprehenders simulate the perceptual state of affairs implied by the sentence, then response times should be faster only when pictures are fully (not partially) congruent with sentence content. Participants (N=200) read sentences like “The sun/the moon is shining onto a horse” followed by pictures depicting mentioned (Experiment 1) or non-mentioned (Experiment 2) objects in either a matching or a mismatching perceptual context (sunny/moony background). Verification times (analyzed with linear-mixed-models) were shorter only when the mentioned (but not non-mentioned) objects matched the perceptual context implied by the sentence. The results are discussed in support of simulation account.