Periodic Reporting for period 1 - INSPIRE (Multilingual sentence processing in real-time)
Reporting period: 2022-03-01 to 2024-02-29
In today's globalized world, the ability to speak and comprehend more than one language is no longer an exception. Many people are bilingual or multilingual speakers. The project INSPIRE (multilINgual Sentence Processing In Real-timE) focused on two factors that have been proposed to influence multilingual sentence processing: language dominance and cross-linguistic influence. The project especially sought to establish how the processing of the first language (L1) is affected by these factors, that is, whether a multilingual speaker’s L1 undergoes attrition. The overarching goal of the project was to shed light on the ongoing debate about how the language systems in a multilingual speaker’s mind/brain interact.
We developed an experimental design that allowed us to test for cross-linguistic influence. We took German as our target language, which has morphological case marking and a relatively free word order. Unlike German, languages such as English and Norwegian lack morphological case marking on noun phrases. In these languages, word order helps listeners identify who did what to whom in a sentence. Hence, L1 speakers of these languages may over-rely on word order when processing sentences in German, due to L1 influence. Conversely, L1 speakers of German may rely more on word order than case when immersed in a Norwegian/English-speaking environment, due to influence from the societal language(s).
A first lab study with German L1 speakers in Germany showed incremental use of morphological case marking during sentence comprehension (Schlenter & Westergaard, 2024, doi: 10.1017/S1366728922000736). Participants in the study were presented with two pictures, one that matched the event as described in an auditorily presented sentence (e.g. secretary bringing minister to general) and one that mismatched (e.g. secretary bringing general to minister). As revealed by their eye movements, L1 speakers in the home country identified the correct picture while the sentence unfolded, and they were able to do so prior to the final case cue when certain conditions were met (e.g. when the previous case cue was perceptually salient). Next, we slightly modified our materials based on our findings and tested German L1 speakers in Norway. The participants in this second study resided either temporarily or more permanently abroad and were regularly exposed to Norwegian and/or English. We found subtle processing differences between this group of speakers and the one in Germany. Especially long-term expats who had spent more than ten years abroad, exhibited a difference in the use of dative relative to accusative case in real-time processing. Our findings align with findings from previous research on heritage speakers of German and research on child L2 attrition that show a specific vulnerability of the dative. Crucially, our findings cannot be explained by incomplete or differential acquisition, as our speakers had fully acquired German as their L1 before moving abroad, and the study thus provides novel evidence for adult L1 attrition. Our findings indicate that, if we take the societal language as a proxy for language dominance, language dominance influences real-time sentence processing.
Finally, we conducted a web-based picture selection task (to reach more participants with a specific language profile) with the same materials as in the second study. In the third study, we systematically investigated whether there is cross-linguistic influence from the L1 to a later-learned language (Ln) by comparing two groups of learners of German: Norwegian L1 and Polish L1. Like German and unlike Norwegian, Polish has morphological case marking and a free word order. For all learners, German was the third or even later acquired language, and all had knowledge of English. Moreover, none of the Ln learners reported regular exposure to/use of another case-marking language. The results showed that Polish learners of German had fewer comprehension difficulties than the Norwegian learners, for example, when selecting the target picture for object-initial sentences in German. As the two groups had a similar language profile, the results indicate that there is cross-linguistic influence from the L1 to a Ln in multilingual sentence processing.
A first lab study with German L1 speakers in Germany showed incremental use of morphological case marking during sentence comprehension (Schlenter & Westergaard, 2024, doi: 10.1017/S1366728922000736). Participants in the study were presented with two pictures, one that matched the event as described in an auditorily presented sentence (e.g. secretary bringing minister to general) and one that mismatched (e.g. secretary bringing general to minister). As revealed by their eye movements, L1 speakers in the home country identified the correct picture while the sentence unfolded, and they were able to do so prior to the final case cue when certain conditions were met (e.g. when the previous case cue was perceptually salient). Next, we slightly modified our materials based on our findings and tested German L1 speakers in Norway. The participants in this second study resided either temporarily or more permanently abroad and were regularly exposed to Norwegian and/or English. We found subtle processing differences between this group of speakers and the one in Germany. Especially long-term expats who had spent more than ten years abroad, exhibited a difference in the use of dative relative to accusative case in real-time processing. Our findings align with findings from previous research on heritage speakers of German and research on child L2 attrition that show a specific vulnerability of the dative. Crucially, our findings cannot be explained by incomplete or differential acquisition, as our speakers had fully acquired German as their L1 before moving abroad, and the study thus provides novel evidence for adult L1 attrition. Our findings indicate that, if we take the societal language as a proxy for language dominance, language dominance influences real-time sentence processing.
Finally, we conducted a web-based picture selection task (to reach more participants with a specific language profile) with the same materials as in the second study. In the third study, we systematically investigated whether there is cross-linguistic influence from the L1 to a later-learned language (Ln) by comparing two groups of learners of German: Norwegian L1 and Polish L1. Like German and unlike Norwegian, Polish has morphological case marking and a free word order. For all learners, German was the third or even later acquired language, and all had knowledge of English. Moreover, none of the Ln learners reported regular exposure to/use of another case-marking language. The results showed that Polish learners of German had fewer comprehension difficulties than the Norwegian learners, for example, when selecting the target picture for object-initial sentences in German. As the two groups had a similar language profile, the results indicate that there is cross-linguistic influence from the L1 to a Ln in multilingual sentence processing.
The project findings extend previous knowledge about sentence processing in German and, crucially, provide new insights into effects of bi-/multilingual experience on sentence processing. Both language dominance and cross-linguistic influence have been confirmed as critical factors in multilingual sentence processing. Moreover, given that bi-/multilingualism is increasingly common, our findings have implications for society. The subtle L1 attrition in adulthood can be of relevance to an increasing number of people who leave their home countries and experience reduced exposure to their L1(s), for example refugees.