Memory consolidation: Surprisingly pervasive effects on language, at multiple levels.
Saturday, October 01st, 2011 [14:20 - 16:00]
SY_16. Memory consolidation: Surprisingly pervasive effects on language, at multiple levels
Dumay, N. 1 & Samuel, A. G. 1, 2, 3
1 Basque Center on Cognition, Brain and Language. San Sebastian, Spain
2 Ikerbasque, Spain
3 Stony Brook University, New York, USA
In this 90-min symposium, four speakers and one discussant examine a range of memory consolidation effects during language learning. Consolidation is not just about making new representations more resistant and easily accessible. Rather, it seems that after the initial exposure to new linguistic stimuli, critical processes take place offline, and in particular during sleep, that substantially change the way we represent the newly acquired information. Kathleen Rastle and Matthew Davis look at the acquisition of new word suffixes and their meaning and ask whether consolidation is required for learners to be able to generalize the use of these new suffixes to other, novel word contexts. Nicolas Dumay and Jeffrey Bowers focus on spoken word form acquisition. By manipulating the exposure and test voices, they examine whether lexical consolidation strips off the surface details of newly learned words, thereby making them more abstract. Gareth Gaskell also investigates lexical learning and tries to understand why certain aspects of this process are immediate while others require overnight consolidation. He reports a critical role for speed of memory retrieval in distinguishing consolidation effects. Howard Nusbaum assesses the beneficial impact of sleep on phonetic learning, and, in particular, in recovering fading perceptual skills. He places phonetic consolidation effects within a broader range of such phenomena, and demonstrates that consolidation plays a powerful role in communication in other species as well. Arthur Samuel will discuss the range of consolidation phenomena, and their implications.
TALKS
SY_16.1 - Does linguistic generalisation require consolidation?: The case of artificial morpheme learning
Rastle, K. 1 & Davis, M. H. 2
1 Department of Psychology, Royal Holloway, University of London, UK
2 MRC Cognition and Brain Sciences Unit, Cambridge, UK
The generalisation of morphological knowledge is at the core of our language abilities. Our experience with a limited set of morphologically-complex words (e.g. unscrew) allows us to abstract knowledge about the components of those words (e.g. un-) that is critical for the interpretation and creation of new words (e.g. untweet). This research examined the role of consolidation in the acquisition of this morphological knowledge using an artificial language paradigm in which undergraduate participants were trained on novel suffixes (e.g. -nept) presented in novel word contexts (e.g. sleepnept). Participants were tested immediately after training or two days after training on a variety of tasks including a speeded auditory repetition test that assessed performance on trained items as well as items comprising an untrained suffix, an untrained stem or both (e.g. sleepnept vs. sleepnule, or sailnept vs. sailnule). Results of the auditory repetition task showed an advantage for trained items immediately after training, as these were repeated significantly faster than items with trained stems but untrained suffixes. Critically, this advantage generalised to untrained items comprising the novel suffixes (e.g. sailnept) but only in the group of participants tested two days after training. This study demonstrates that participants can acquire morphological representations in a laboratory context that are sufficiently robust to generalise to the interpretation of novel exemplars. It also suggests that this process of linguistic generalisation may require a period of overnight consolidation consistent with complementary systems theories of lexical learning.
SY_16.2 - Does lexical consolidation make words more abstract?
Dumay, N. 1 & Bowers, J. F. 2
1 Basque Center on Cognition, Brain and Language. San Sebastian, Spain
2 School of Experimental Psychology, University of Bristol, UK
Learning new words involves consolidation. After one night's sleep, not only is explicit knowledge about the novel words enhanced, but the new words also now compete with similar-sounding existing words (Dumay & Gaskell, 2007) during word recognition. The present study assessed whether lexical consolidation strips off surface details of newly learned words, producing more abstract representations. We manipulated the speaker's voice between exposure and test. Participants learnt one set of novel competitors (such as 'shadowks' for 'shadow') seven days before the test, and another set immediately before the test. Each word was learnt in a male or a female voice, and was tested in either the same or the other voice. Cued recall ('shadow--?') and phoneme monitoring (Is there an /s/ in 'shadowks'?) showed stronger memory (i.e., more accurate/faster responses) for the seven-day old items and, if anything, an enhanced voice effect (i.e., better performance in the same voice condition) after seven days. Crucially, our most indirect measure of lexical competition showed that only the seven-day old items (as expected) engaged in lexical competition, but only when the input preserved the voice in which they had been encoded. These findings indicate that consolidation does not make word representations more abstract: voice specific details do not just survive lexical consolidation; they are enhanced by it.
SY_16.3 - Learning new vocabulary: Why the delay?
Gaskell, G.
Department of Psychology, University of York, UK
Evidence is accumulating that the acquisition and retention of novel words involves an extended process of consolidation, with some immediately observable lexical effects, and some more protracted sleep-associated characteristics. This profile of acquisition suggests that language learning is constrained by more general properties of memory, and is consistent in broad terms with a complementary systems approach. However, it is less clear what defines a delayed property and why. Here, I will discuss two possible models. In the first, immediate and delayed components of a new word representation are distinguished by the level of integration between recent and remote memories of lexical items. In the second, the key distinguishing feature is the speed of retrieval of the recently learned lexical representation. Recent evidence from word learning studies of both form and meaning will be evaluated against these two models, and I will argue that the speed-of-retrieval model provides the best fit to the existing data.
SY_16.4 - Sleep consolidation of phonetic learning
Nusbaum, H.
Center for Cognitive and Social Neuroscience, Department of Psychology, The University of Chicago, Chicago, IL USA
Language learning is marked by two broad characteristics: (1) Learning generalizes to new patterns and situations, and (2) learning is robust, showing little performance variability over time. There is substantial plasticity in adult learning phonetic patterns, and there is evidence of both of these characteristics. Recently we investigated how sleep contributes to the long-term stability of phonetic learning through consolidation (Fenn et al., 2003). The basic paradigm compares posttest to pretest recognition performance after a single training session. By varying duration of a retention interval and inclusion of sleep (immediate, 12 hours awake, 12 hours with sleep), we can assess the effect of sleep on consolidating phonetic learning. Moreover, by examining training and testing at different points in a day, we can control for circadian contributions. We found that after a 12 hour waking retention interval, there is a significant loss of performance compared to a posttest right after training. However, sleep restores this loss of performance. Moreover, sleep after training prevents subsequent loss the next day. We have subsequently demonstrated that this sleep consolidation is not specific to phonetic learning; sleep consolidates generalization learning in a complex sensorimotor task using videogames (Brawn et al., 2008). Using a nap paradigm, we have investigated how much sleep is necessary to consolidate learning. We have also examined how sleep consolidation may differ in rote and generalized learning of phonetic patterns. We have also developed an animal model using an auditory classification task for vocal communication (Brawn et al., 2010). This has provided the first behavioral evidence for sleep consolidation in non-human animals, showing a very similar pattern of consolidation to humans. Using this animal model, we have begun to investigate the nature of the loss of performance over a waking retention interval and how sleep may restore performance.