This collection offers various psycholinguistic perspectives on the link between the processes of speech comprehension and production. Some authors in this collection approach the topic by reviewing how well existing models deal with the comprehension-production relationship; others propose new approaches, including ones that show compatibility with evidence from neuroimaging studies. The final two of the nine papers address the topic from a SLA standpoint.

Dell and Gordon use an interactive activation model of lexical access to account for the differential effects of phonological neighborhood density (number of words phonologically similar to the target) on recognition and production. For the word recognition process, the set of potential candidates is phonologically based, accounting for the inhibitory effect of high neighborhood density. These same neighbors have a facilitative effect on production because competitors in production are semantically versus phonologically related words.

McQueen, Dahan, and Cutler suggest that parallel activation models are well-suited to the continuous fine-grained processing that maps the acoustic-phonetic input to stored memory representations in word comprehension. However, they note that the sequencing of information in production through discrete stages of processing is more categorical in nature.

Zwitserlood argues for a separate representation of morphological information in the lexical system and suggests that although separate phonological and phonetic representations might be involved in comprehension and production, the syntactic and proposed morphemic levels may be shared by these two processes.

Roelofs reviews computational models of word recognition and production and concludes that cohort competitors (e.g., captain and captive) and rhyme competitors (e.g., dollar and collar) have different effects on word recognition and production tasks. To account for the differential influence of serial order on lexical and sublexical levels in these tasks, Roelofs proposes distinct but closely linked feedforward-only systems for word production and recognition. This is compatible with neuroimaging studies showing that these processes involve separate regions within Wernicke's area.

Goldstein and Fowler adopt the perspective of articulatory phonology, which captures production as a coordination of gestures into more elaborate temporally overlapping structures, accounting for the context-dependence of speech. They argue that these gestures as vocal events are also the targets of listener perception.

Guenther proposes a neural network model of speech production and motor acquisition, encompassing the speech-planning process through to the level of muscle commands. This is also a model of the neural representations that underlie production, and it suggests a role for auditory cortical areas in both production and perception.

Van Turennout, Schmitt, and Hagoort focus on how data from event-related potentials can help to determine the sequence and time course in single-word processing of the retrieval of semantic, syntactic, and word form (i.e., phonological or orthographic) information by speakers and listeners (or readers). The retrieval sequence of semantic—syntactic—phonological form holds for tasks such as picture naming. For comprehension, access to phonological form takes precedence, with evidence that semantic information is retrieved before syntactic. This sequence is compatible with different architectures of the cognitive system.

Sebastián-Gallés and Kroll review comprehension and production studies on phonological acquisition and organization by bilingual infants and adults. They define bilinguals as “individuals who use more than one language on a regular basis” (pp. 279–280). Evidence strongly suggests that word activation is based on phonological similarity to the target, regardless of the language. In production, the phonology of the translation equivalent of the target is activated in addition to the target's phonological neighbors.

Finally, Flege discusses the acquisition of phonetic categories by second language (L2) learners. The speech learning model assumes that adults retain the ability to develop new L2 phonetic categories; however, the success of the process declines with age and is compromised by first language phonetic categories, which may assimilate similar L2 speech sounds. He proposes that the accuracy of L2 perception constrains that of L2 production. This contrasts with results of several studies reporting that production of /r/ and /l/ by some Japanese speakers was more accurate than their perception of these sounds (e.g., Sheldon & Strange, 1982). Additionally, considerable individual variation in perception and production learning has been documented (e.g., Bradlow, Pisoni, Akahane-Yamada, & Tohkura, 1997). It is unfortunate that this latter work on the effects of perceptual training on production improvement is mentioned only briefly (p. 345), given the main thrust of the volume on the perception-production relationship.

In sum, this collection contributes substantially to an understanding of lexical organization and the challenges for models addressing the relationship between comprehension and production. All papers will be of interest to speech researchers and are recommended reading for graduate students in the field.