Incremental learning in word production: Tracing the fate of non-selected alternative picture names.

Abstract

Picture naming studies have shown that alternative picture names become phonologically coactivated even when they are eventually not produced (e.g., dog when poodle is produced and vice versa). The authors investigated whether this pattern is shaped by recent experience. Specifically, they tested whether the phonological coactivation of an alternative name is attenuated, when speakers consistently only use one particular name in a large number of naming episodes. In 3 picture-word interference experiments, the authors measured the phonological coactivation of basic-level alternative names during subordinate-level naming (Experiments 1 and 3) and of subordinate-level alternative names during basic-level naming (Experiment 2). They measured the coactivation of the alternative name at different points in time with distractor words that were phonologically related or unrelated to that name. If the pattern of lexical activation is shaped by previous naming episodes and the phonological coactivation of the nonproduced alternative name decreases, interference from related distractors should be reduced in the course of the experiment. Contrary to this prediction, the interference effect from distractors phonologically related to the alternative name remained stable. This was also true when participants were not familiarized with the pictures, more naming episodes were implemented, and a consolidation phase was introduced (Experiment 3). Overall, these results indicate some limitation of incremental learning in word production. (PsycINFO Database Record (c) 2018 APA, all rights reserved).

Pragmatic constraints do not prevent the co-activation of alternative names: evidence from sequential naming tasks with one and two speakers

“…Following these previous studies, we interpret this phonological interference effect as an indicator of phonological co-activation of alternative picture names (Jescheniak et al, 2005;2017;Kurtz et al, 2018). A phonological locus is supported by time-course estimates and theoretical considerations (Jescheniak & Schriefers, 1998; see also Jescheniak, Hahne, Hoffmann, & Wagner, 2006).…”

“…Latencies deviating more than 2 SD from a participant's and an item's mean per experimental condition were classified as outliers and excluded from further analysis as well (158 observations, 1.9%). This outlier criterion was determined prior to any data analysis and is the same one we used in previous related studies, Jescheniak et al, 2005;2017;Kurtz et al, 2018).…”

“…The distractor manipulation (alternative-related vs. unrelated) is the same as in previous related studies. In neutral contexts, distractors phonologically related to an alternative picture name interfere more strongly with the naming response than unrelated distractors (Jescheniak et al, 2005;2017;Kurtz et al, 2018). This effect can be explained as follows.…”

“…data sets with varying degrees of reduction of the phonological interference effect in the constraining context compared to the neutral context. The expected phonological interference effect in the neutral context was based on a parameter estimate of this effect across all available experiments which had tested for this effect using the same material set (i.e., the present data and the data of Jescheniak et al, 2017, andKurtz et al, 2018). Expected variances across items and participants were based on random effect estimates across these data sets.…”

“…Recent studies have shown that even though the basic-level name is not produced in such a situation, it becomes nevertheless lexically co-activated up to a phonological level (Jescheniak, Hantsch, & Schriefers, 2005;Jescheniak et al, 2017;Kurtz, Schriefers, Mädebach, & Jescheniak, 2018; see also Jescheniak & Schriefers, 1998;Peterson & Savoy, 1998, for related evidence on synonyms). For example, a distractor word phonologically related to the target picture's basic-level name (e.g., "finger" related to "fish", when "shark" is the target) slows down subordinate-level picture naming compared to an unrelated distractor (e.g., "book").…”