The influence of age on the time course of word preparation in multiword utterances

Abstract

Sentence production requires speakers to co-ordinate the preparation of words so that they are ready for articulation when they are needed. Ageing appears to influence both the speed and likelihood of successful word retrieval. We examine how age differences in word production might influence the production of larger units of speech such as sentences. Speakers described displays containing three objects of systematically varied naming difficulty. The latency, duration, content, and fluency of speech in addition to its co-ordination with eye movements indicated that both young and older adults prepared their words immediately before uttering them. As a consequence, older adults were also significantly less fluent in their utterances than were younger adults.

Acknowledgement

This research was supported by National Institute on Aging Grant RO1 AG17024. We would like to acknowledge Amit Mookerjee for assistance in software development, to Ellen Hamilton, Edna Chavira, and Nadine de Lassus for the substantial effort involved in data collection, and to Justin Storbeck, Jordan Wang, and Russell Smith for transcriptions and speech measurements. We would also like to thank Elisa Lawler and Tracy Mitzner for comments on an earlier version of this manuscript and Demetria Mitchell for assistance in manuscript preparation. Thanks also to the Cognitive Aging community at Georgia Tech for invaluable discussion.

Notes

¹The term codability is used to highlight the fact that the objects ate readily identified but differ in the number of names they elicit. Moderate name agreement alone can be due to the same object having multiple appropriate names or being very difficult to identify.

²An interaction between codability and frequency would be predicted by most interactive activation theories under these assumptions. However, it is unclear under which conditions activation from earlier levels of representation may facilitate the retrieval of phonological information in Node Structure Theory. For example, Rastle and Burke (1996) predicted and found that earlier semantic processing of words did not affect TOT rates, although one would expect extra activation of semantic representations to be passed on to phonological ones.

³The degree of incrementality observed in these studies differs from that argued for in other similar non-eye tracking experiments such as Smith and Wheeldon (1999) and Martin, Miller, and Vu (2004). Those studies concluded that entire subject noun phrases were encoded prior to speech onset based on faster onset latencies for sentences such as “The A moves above the B and C” relative to “The A and B move above the C”. While such results could be considered evidence that two nouns were encoded before speech onset in the latter sentence but not in the former, the small latency difference (under 100 ms) and the lack of any evidence for word specific processing of B object names suggest instead that some other aspect of preparing complex subject noun phrases delayed speech onset slightly or that preparation of B was begun but not completed (as in Griffin, 2001, 2003).

⁴Transcriptions and soundfiles were also run through a forced alignment program, Fasttalk™, to measure noun onsets. Fasttalk yielded noun onsets with confidence values greater than zero for 81% of the first nouns and 89% of critical nouns in analysed trials. The agreement between the forced alignments and hand measurements was generally high with a mean difference of 19 ms for A names and 28 ms for critical nouns. Only the hand measurements are used here.

⁵Although first object names were matched for length pre-experimentally, speakers tended to replace low frequency names with longer words than high-frequency names, resulting in longer low frequency names. These longer first names allowed more time after the onset of speech for speakers to prepare second nouns (see Griffin, 2003, for a full account of this reversed length effect). The presence of this effect suggests that speakers were making an effort to produce object names fluently and were willing to briefly buffer short first object names.

⁶Analyses were also performed excluding trials with non-targets produced for the first object name to ensure the results were not due to differences in the words used and their lengths. The results were the same as those including those trials.

Additional information

Notes on contributors

Daniel H. Spieler

Correspondence should be addressed to either author at the School of Psychology, 654 Cherry Street, Georgia Institute of Technology, Atlanta, GA 30332-0170, USA. [email protected] or [email protected]