Don’t Underestimate the Benefits of Being Misunderstood

Abstract

Being a nonnative speaker of a language poses challenges. Individuals often feel embarrassed by the errors they make when talking in their second language. However, here we report an advantage of being a nonnative speaker: Native speakers give foreign-accented speakers the benefit of the doubt when interpreting their utterances; as a result, apparently implausible utterances are more likely to be interpreted in a plausible way when delivered in a foreign than in a native accent. Across three replicated experiments, we demonstrated that native English speakers are more likely to interpret implausible utterances, such as "the mother gave the candle the daughter," as similar plausible utterances ("the mother gave the candle to the daughter") when the speaker has a foreign accent. This result follows from the general model of language interpretation in a noisy channel, under the hypothesis that listeners assume a higher error rate in foreign-accented than in nonaccented speech.

Speaker-specific processing of anomalous utterances

“…Individuals made more nonliteral plural interpretations of sentences containing plural verbs and plural nouns (vs those with singular nouns and singular verbs) and more non-literal "unblended" interpretations of plausible implicit negation blends (vs control stimuli). This is consistent with the finding that comprehenders interpret utterances following what is inferred to be likely rather than what has literally been observed, as outlined by noisy channel and good-enough processing frameworks (e.g., Christianson et al, 2001Christianson et al, , 2010Ferreira, 2003;Gibson et al, 2013;Gibson et al, 2017;Frazier & Clifton, 2015;Levy, 2008). The agreement findings replicate previous work (Patson & Husband, 2016), while the without-blend findings demonstrate that responses to comprehension questions mirror semantic judgments (Frazier & Clifton, 2015).…”

“…To the extent that a comprehender attributes the anomaly to their own miscomprehension, they may be more likely to infer that the head number should be changed; to the extent that a comprehender attributes the anomaly to the speaker's misproduction, they may be more likely to infer that the head number was produced as intended. The dissociation between repairs to be made for comprehension-and production-centred inferences sets these constructions apart from those used in previous work (e.g., "The mother gave the candle the daughter"; e.g., Gibson et al, 2013;Gibson et al, 2017), where there is only one parsimonious repair to be made (add to). This allows us to examine the unique contributions of reader miscomprehension and speaker misproduction to non-literal inferences.…”

Quasi-compositional mapping from form to meaning: a neural network-based approach to capturing neural responses during human language comprehension

“…In line with previous proposals of a temporary 'semantic illusion' [58] and with independent evidence suggesting influences of plausibility on comprehension (e.g. 25% of college students report understanding the sentence 'The dog was bitten by the man' as indicating that the dog was the agent of the biting action [56,57,59,60]), the model's representations are influenced by plausibility in addition to syntactic cues, so that it can end up in a state where the interpretation remains dominated by its experience with the typical roles played by objects in events (e.g. eggs being eaten rather than eating something).…”

“…In general, it seems beneficial to rely on all possible sources of information rather than giving overriding importance to a single consideration. In line with this view, rates of plausibility based interpretations are higher when listening to speakers with a foreign accent, suggesting that such interpretations are not a failure of the system, but rather result from a Bayes optimal process taking into account all available cues to best estimate the intended meaning, integrating noisy evidence and semantic priors [57]. These considerations play into an important contrast between the SG model and an alternative model by Brouwer et al [39], which is grounded in the classical view that syntax necessarily has a decisive role in interpretation.…”

“…Ingredients of a solution to this issue might include prioritization of specific cues via attention-dependent signals as proposed in a model of cognitive control in the Stroop task [67]. For example, the aim to understand a foreigner's intended meaning despite possible syntactic errors might result in an up-weighing of semantic cues [57]. By contrast, the aim to enjoy fairy tales might result in an up-weighing of syntactic cues to determine the syntactically indicated meaning despite its possible real-world implausibility.…”