Factors affecting cross-language activation and language mixing in bilingual aphasia: A case study

ABSTRACT

Background

Some bilinguals with aphasia tend to mix languages within a single utterance. Two opposing views attribute this to difficulty with either language control or word retrieval.

Aims

This study investigated the influence of factors that increase activation of the non-target language on the occurrence of language mixing errors. This increased activation predicts more language mixing errors if there is a language control issue, but not if they stem from word retrieval difficulties.

Methods and procedures

A picture naming experiment was conducted with a bilingual individual with aphasia who showed language mixing. We investigated the influence of four factors likely to influence activation of representations of the non-target language on response accuracy, response latency and the occurrence of language selection errors: language, language mode, task, and phonological overlap between the target word and its translation equivalent.

Outcomes and results

The increased activation of the non-target language induced by language mode, task and phonological overlap with the translation equivalent did not lead to an increase in language selection errors when compared to correct responses. This is despite the fact that these factors affected accuracy and response latency, in the direction that is expected in unimpaired bilingual performance.

Conclusions

Results were not consistent with a disruption of the cognitive control needed to respond in the intended language. Instead, they highlight that language mixing in this individual, rather than being “pathological”, is instead used as a strategy to potentially improve communication when lexical retrieval difficulties occur. Language mixing behaviours in aphasia may not be due to issues of control and have a communicative value that should be recognised.

KEYWORDS:

• bilingual aphasia
• language mixing
• language mode
• cognates

Supplemental data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/02687038.2022.2081960

Acknowledgements

During the preparation of this article, Solène Hameau was funded by an Australian Research Council Discovery Program grant (DP190101490).

We thank CA for his enthusiastic participation in this study. We also thank Jaime Undurraga for valuable assistance with testing in Spanish, Macarena Moreno for help with stimulus preparation, assessment and response transcription, Josué Pino Castillo and Alejandra Romano for assistance in response transcription.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

^1. It is important to note that while these steps were to ensure strong activation of both languages, the resulting communication situation does not correspond to a bilingual mode in the sense Grosjean describes it, in that the participant was still explicitly required to respond in a particular language at a given time, and not the other language. A true bilingual mode would have meant that the participant was “allowed” to freely use either language, which was not the case here.

^2. The same analysis was performed on untransformed response times and showed the same pattern of significance, except for the interaction between language and phonological distance, which was now just under the significance threshold (p = .075).

^3. Analyses on untransformed response times led to the same pattern of significance.

^4. In an omission, no word form is sufficiently activated to reach a level that allows its selection. In that sense, a factor that makes it more likely to select a word in the other language than no word at all as is the case with an omission, can be seen as “facilitatory”. Another source of evidence showing that a language error can be seen as facilitated retrieval as compared to an omission is the effect of frequency that was observed in CA: words were more likely to result in a language error than in an omission if they were more frequent (hence, “easier words” were more likely to result in a language error than in an omission).

Additional information

Funding

This work was supported by the Australian Research Council [DP190101490].