Brief Lecture Notes for Unit 5
Because this unit is about sensation and perception, and because I don't want to give myself a case of carpal tunnel syndrome, I'm going to hyperlink a few passages that provide basics about this topic, from PSY 202. The material you will need to know is in red font, like this... you may safely ignore the other material on those pages.
General overview of sensation and perception
The "ladder of learning" as related to methodologies in sensation and perception
In Unit 2, we discussed some reasons why it's more difficult in general to conduct research with children than with adults. These problems are even greater when it comes to perceptual research... particularly, perceptual research with infants (which is the sort of perceptual research we need to address such matters as the nativist-empiricist controversy). Here are some of the reasons why:
1. The problem of self report. Of course, prelinguistic children can't give us any sort of direct self-report at all; they can provide some rudimentary forms of indirect self-report by means of operant conditioning (see the second hyperlink above). But they can't provide any sophisticated or nuanced feedback; at best, they can offer "yes or no" type responses. And even then there is much more "noise" in the response data than we would expect with older child or adult respondents.
2. The problem of shifting mental states. In all research contexts... come to think of it, in pretty much any social context whatever... mental states are a problem. (For instance, teachers have to worry about boring their students to death, because effective learning stops when students' minds start to wander.) But we have to be much more concerned about this issue with young children and infants, because their attention span is so much shorter; because the outward signs of inattention may be less obvious; because (due to differences in the sleep-waking cycle that are neurologically based) sudden shifts in mental states (e.g., from alert wakefulness to the early stages of sleep) are so much more prevalent; because they are more easily bored (habituated) in the presence of stimuli that are incomprehensible or unmotivating, which includes a much larger class of stimuli than would be the case for adults.
As noted above, perceptual research with infants usually utilizes one of the elementary forms of learning (habituation, respondent conditioning, or operant conditioning) as a means of circumventing the self-report problem.
Many of the studies that result from these methodologies point strongly in the direction of the nativist view of perception, though this point is hotly debated (mostly by long-standing empiricists who see their job security at risk). For instance, it appears that very early in life, infants respond differentially to the human face... much more so than to equally complex visual stimuli that are not facelike. They are able early on to discriminate between rightly constructed faces and "jigsaw" faces that have the individual components misplaced. This probably points to a hard-wired, genetically programmed capacity to respond to social stimuli that (as we'll see in Unit 6) plays a very significant role in infant attachment and mother-infant bonding.
Similarly, key perceptual properties like depth perception seem more or less hard-wired, as shown by means of Bower's experiment, as well as similar experiments such as the visual cliff (I'm not going to describe each of these experiments in detail because your textbook does a fine job of that, and because you need to be attending lecture, anyway).
A particularly intriguing instance of apparent genetic hard-wiring (one that poses some interesting explanatory challenges) has to do with the seemingly pre-programmed ability of infants to discriminate between similar phonemes (speech sounds). By means of an electronic speech synthesizer, we can take two existing phonemes like /p/ and /b/, which differ only in a single attribute (/p/ is "unvoiced", while /b/ is "voiced"), and construct artificial hybrids or blends that lie partway in between the two, e.g., a hybrid that is 70% /p/-like, 30% /b/-like. In this way we can turn something that is normally dichotomous (either-or) into an auditory continuum. We can then compare infants' response to this "speech continuum" to a control condition (a "musical continuum" of musical tones -- notes of different auditory frequencies lying between middle C and C-sharp, for instance).
Using habituation as a methodology, we first habituate the infants to a particular sound (/p/ in the one case, middle C in the other case). We then present a randomized array of other sounds (phonemes or musical tones) and measure the infant's degree of response. In the case of the musical tones, we get a linear response curve (I don't have the software to put these curves in this Web site, you'll have to attend class), as we would normally expect. But with the languagelike sounds, we get a sharp response boundary, as if the child is hard-wired to maximize the difference between similar sounds. Without this ability, no child would ever learn language... so the interesting question can be raised, how did this property get there in the first place? (It would have no adaptive significance in a nonlinguistic environment, of course.)
Study Guide
1. How do sensation and perception differ? What are three reasons why contemporary psychologists do not draw a sharp distinction between sensation and perception? Explain.
2. What is meant by the "ladder of learning"? Explain how the techniques of habituation, respondent conditioning, and operant conditioning can be used to conduct perceptual research with infants.
3. What are two key problems in conducting perceptual research with infants? Discuss. How can each be addressed or circumvented?
4. Summarize and critique experimental evidence relevant to the nativist-empiricist controversy in perception.