![Sample our Behavioral Sciences journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/110801/TOC-Subject-Sample-2021-Behavioral-Sciences.jpg"})
ABSTRACT
Despite early evidence that grammatical gender is retrieved prior to phonology in lexical access, more recent studies demonstrating task effects and non-converging evidence raise doubts about the extent to which this is a general feature of the language production system. We employed the dual-choice go/no-go paradigm with event-related potentials (ERPs) in order to further clarify the time course of retrieval of grammatical gender and phonology. Specifically, we examined how task order influences the relative timing with which these features are retrieved. Results find no clear evidence that grammatical gender is retrieved prior to phonology in a serial manner. Instead, the relative timing with which these features are retrieved is subject to task order, suggesting that prior estimates of lexical access obtained with this paradigm may be confounded by task effects. Overall, our result support parallel access models of feature retrieval during lexical access and suggest that attentional biases may modulate retrieval.
Acknowledgements
We thank Nyssa Bulkes, Jack Dempsey, Lorenzo Grego, Chase Krebs, Yisi Liu, Maha Tolba, Rebekah Seyfert and Fiona Weingartner for help with data collection. Thanks to Silvina Montrul and two anonymous reviewers for comments on an earlier version of this manuscript. Any remaining errors are our own.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. It is worth noting, though, that Van Turennout et al. (Citation1998) measured the difference between retrieval of grammatical gender and phonology as the difference between the latency at which the go and no-go LRPs diverged for hand = gender and the latency at which the no-go LRP onset (i.e. divergence – no-go onset). Other studies have used the onset of the go LRP rather than the no-go LRP to calculate this difference (i.e. divergence – go onset), in which case the difference in Van Turennout et al.’s study is 20 ms rather than 40 ms, which is much closer to the difference found by Camen et al. (Citation2010).
2. Half of the participants participated in both experiments.
3. Note that it is not possible to compare onset latencies for go LRPs to phonology, as these were based on word-final phones in Van Turennout et al. (Citation1997), which take longer to retrieve than word-initial phones (see Indefrey, Citation2011 for an overview).
4. Removing left-handed participants from the analyses did not qualitatively change the pattern of results seen in the full participant group. Data from all participants are thus reported. Results of analyses excluding left-handers are reported in the supplementary materials in Tables S1–S13 and Figures S12–S16. Note that for the N200 analyses, there is a small change in statistical significance when excluding left-handers, despite the overall pattern of peak latencies remaining the same (Table S9). We attribute this change to power loss associated with the smaller sample excluding left-handers.
5. Prior to artefact rejection, trials with the items das Kissen and das Bügelbrett were removed for the Austrian participants due to the use of different lexical items for these objects in Austrian dialects (der Polster and der Bügelladen, respectively). These are not counted as lost trials.
6. Three nouns from the dual-choice go/no-go task were repeated in the gender decision task. These did not have any effects on the results, and are therefore included in all statistical analyses of the gender decision data. Results of the analyses excluding these items from the gender decision task are reported in the supplementary materials in Tables S14–S17.
