Cue integration with categories: Weighting acoustic cues in speech using unsupervised learning and distributional statistics

During speech perception, listeners make judgments about the phonological category of sounds by taking advantage of multiple acoustic cues for each phonological contrast. Perceptual experiments have shown that listeners weight these cues differently. How do listeners weight and combine acoustic cues to arrive at an overall estimate of the category for a speech sound? Here, we present several simulations using a mixture of Gaussians models that learn cue weights and combine cues on the basis of their distributional statistics. We show that a cue‐weighting metric in which cues receive weight as a function of their reliability at distinguishing phonological categories provides a good fit to the perceptual data obtained from human listeners, but only when these weights emerge through the dynamics of learning. These results suggest that cue weights can be readily extracted from the speech signal through unsupervised learning processes.     

The End(s) of Philosophy: Rhetoric, Therapy and Wittgenstein's Pyrrhonism

In Culture and Value Wittgenstein remarks: ‘Thoughts that are at peace. That's what someone who philosophizes yearns for’. The desire for such conceptual tranquillity is a recurrent theme in Wittgenstein's work, and especially in his later ‘grammatical‐therapeutic’ philosophy. Some commentators (notably Rush Rhees and C. G. Luckhardt) have cautioned that emphasising this facet of Wittgenstein's work ‘trivialises’ philosophy – something which is at odds with Wittgenstein's own philosophical ‘seriousness’ (in particular his insistence that philosophy demands that one ‘Go the bloody hard way’). Drawing on a number of correlations between Wittgenstein's conception of philosophy and that of the Pyrrhonian Sceptics, in this paper I defend a strong ‘therapeutic’ reading of Wittgenstein, and show how this can be maintained without ‘trivialising’ philosophy.     

Concurrent Validity of the Schwartz Outcome Scale With a Chemically Dependent Population

Clients (N = 151) completed the Schwartz Outcome Scale (SOS‐10; M. A. Blais et al., 1999) and the Substance Abuse Subtle Screening Inventory‐3 (F. G. Miller & L. E. Lazowski, 1999). The results show a negative relationship between substance use symptoms and SOS‐10 scores and a positive relationship between defensiveness and SOS‐10 scores.     

Human variable ratio performance

The generality of the mathematical principles of reinforcement (MPR) was tested with humans. In Experiment 1A, participants’ mouse clicks were reinforced according to a series of variable-ratio (VR) values. In one condition, a situated task was used (searching for treasure on a map presented via a PC monitor); in the other, abstract version of the task, participants clicked on a white screen. Under the “map” condition, response rates increased with increasing ratio value before decreasing with further ratio increases; under the “no-map” condition, response rates decreased linearly from a peak at the smallest ratio. In Experiment 1B, the pattern of responding was confirmed under the “map” conditions using a different set of ratio values. In Experiment 2, reinforcer magnitude was manipulated using the “map” context. Response rates did not differ significantly when large rather than small cash reinforcers were delivered according to a VR 30 but were significantly higher when large rather than small cash reinforcers were delivered according to a VR 120. Together these experiments offer qualified support for extending MPR to the behavior of human participants.     

Eyetracking and selective attention in category learning

An eyetracking version of the classic Shepard, Hovland, and Jenkins (1961) experiment was conducted. Forty years of research has assumed that category learning often involves learning to selectively attend to only those stimulus dimensions useful for classification. We confirmed that participants learned to allocate their attention optimally. We also found that learners tend to fixate all stimulus dimensions early in learning. This result obtained despite evidence that participants were also testing one-dimensional rules during this period. Finally, the restriction of eye movements to only relevant dimensions tended to occur only after errors were largely (or completely) eliminated. We interpret these findings as consistent with multiple-systems theories of learning which maximize information input in order to maximize the number of learning modules involved, and which focus solely on relevant information only after one module has solved the learning problem.     

Infants are sensitive to within-category variation in speech perception

Previous research on speech perception in both adults and infants has supported the view that consonants are perceived categorically; that is, listeners are relatively insensitive to variation below the level of the phoneme. More recent work, on the other hand, has shown adults to be systematically sensitive to within category variation [McMurray, B., Tanenhaus, M., & Aslin, R. (2002). Gradient effects of within-category phonetic variation on lexical access, Cognition, 86 (2), B33-B42.]. Additionally, recent evidence suggests that infants are capable of using within-category variation to segment speech and to learn phonetic categories. Here we report two studies of 8-month-old infants, using the head-turn preference procedure, that examine more directly infants' sensitivity to within-category variation. Infants were exposed to 80 repetitions of words beginning with either /b/ or /p/. After exposure, listening times to tokens of the same category with small variations in VOT were significantly different than to both the originally exposed tokens and to the cross-category-boundary competitors. Thus infants, like adults, show systematic sensitivity to fine-grained, within-category detail in speech perception.     

Equivalence relations, contextual control, and naming

This paper reports two experiments that investigated the role of verbal behavior in the emergence and generalization of contextually controlled equivalence classes. During both experiments, participants were trained with two different combinations of the same easily nameable, yet formally unrelated, pictorial stimuli. Match-to-sample baselines for eight four-member classes were established under the contextual control of two colors. In the presence of one color, conditional relations were established between stimuli whose normative names rhymed. In the presence of the other color, conditional relations were established between stimuli whose normative names did not rhyme. Although, during Experiment 1, all participants demonstrated equivalence classes involving rhyming stimuli, none demonstrated the formation of nonrhyme equivalence classes. To investigate this finding, Experiment 2 evaluated whether participants would demonstrate both rhyme and nonrhyme equivalence classes given more extensive exposure to the experimental contingencies. All participants demonstrated contextually controlled rhyme and nonrhyme equivalence classes, although rhyme classes were demonstrated with greater facility than nonrhyme classes. Results indicate that visual stimuli are named, that verbal bases for stimulus classification can affect the emergence of contextually controlled equivalence classes, and that untrained contextually controlled conditional discriminations involving novel stimuli can emerge on the basis of participants' verbal behavior.