JOURNAL OF MEMORY AND LANGUAGE 37, 463–480 (1997) ARTICLE NO. ML972535 Similarity Mapping in Spoken Word Recognition Cynthia M. Connine, Debra Titone, Thomas Deelman, and Dawn Blasko State University of New York at Binghamton Five phoneme monitoring experiments are reported that investigate the relationship between the degree of lexical activation (based on similarity of the input to a real word) and phoneme perception. Experiment 1 showed that phoneme monitoring detection times increased as similarity of the carrier to a real word decreased. Experiment 2 replicated these results with a set of two syllable stimuli. Experiment 4 extended the results of Experiments 1 and 2 to a new phoneme. Two additional control experiments (Experiments 3 and 5) were conducted with truncated stimuli where lexical contributions were removed. The results are discussed in terms of the architectural relationship between the acoustic-phonetic input, form-based lexical levels, and semantic levels of representation.  1997 Academic Press It has long been observed that information tory signal and a lexical representation re- in the speech signal is variable. The earliest quires an unambiguous, canonical input. In spectrograms created at Haskins Laboratories general, this research reveals the remarkable in the 1950s attested to many of the seemingly tolerance of spoken word recognition pro- unpredictable features of the acoustic-pho- cesses to the presence of erroneous or ambigu- netic information that underlies speech per- ous signals. ception. These observations were the basis for In a series of papers, we have begun explor- much of the early research that investigated ing the nature of the mapping process between how individual speech sounds were recovered the speech signal and lexical representations from the speech signal. More recently, re- (Connine, 1994; Connine, Blasko, & Titone, searchers have begun investigating how the 1993; Connine, Blasko, & Wang, 1994; Deel- speech signal is mapped onto lexical represen- man & Connine, 1996). The experiments we tations for spoken word recognition (see have conducted suggest a number of general Miller & Eimas, 1995 for a review). A major properties of the mapping process that under- focus of this work has been the degree to lies spoken word recognition. We have fo- which the mapping process between the audi- cussed on vertical similarity, that is, the notion that lexical activation is proportional to the The research was supported by NIDCD Grants R29 degree there is a successful mapping of the NS26587 and R01 DC02134 to the first author. Additional input to a lexical representation (vertical simi- support was provided by the Center for Cognitive and larity mapping). One general property of this Psycholinguistic Science. Portions of this research were mapping process is that small distortions in reported at the 1992 meeting of the Midwestern Psycho- logical Association, the 1993 meeting of the Psychonom- the signal impede but do not block activation ics Society and at a symposium on auditory word recogni- of the intended word. In fact, spoken word tion at the Sixth Annual (1993) CUNY conference on recognition appears to be remarkably tolerant Sentence Processing. The first author also acknowledges of distortions in the speech signal. A second the support of the Max-Planck Institute for Psycholinguis- general property is that lexical activation is tics in Nijmegen. Debra Titone is now at Brandeis Univer- sity and Dawn Blasko is now at Pennsylvania State Uni- intimately tied to the degree of match (or mis- versity, Erie. match) between a stimulus and a particular Address reprint requests and correspondence to Cynthia lexical representation. There are a number of Connine, State University of New York at Binghamton, lines of evidence that support this general con- Department of Psychology, Binghamton, NY 13901. E- mail: connine@binghamton.edu. clusion. Connine et al. (1994) showed that 463 0749-596X/97 $25.00 Copyright  1997 by Academic Press All rights of reproduction in any form reserved.