I
EXPLORATIONS IN MELODY AND TONAL FRAMEWORK
W. Jay Dowling
University of Texas at Dallas
As the Gestalt psychologists began reminding us three quarters of a century ago, a melody is not just a collection of independent notes, but an integrated whole whose parts are inseparably related in an organic structure. A change in one part of the melody inevitably affects the way the rest of the melody is perceived. Perceiving a melody invokes a context of actual and possible events that adds meaning to what is heard. Melodies are like sentences in these two respects: Sentences are integrated wholes and not just strings of words; and sentences rely for their meaning on a surrounding context of language structure and usage. Melodies depend on the structural context provided by a musical style. The studies reported in the following chapters involved listeners who grew up in a western European tradition of tonal music, and so focus on that tradition in assessing the ways in which melodies are understood and remembered. Just as with the understanding of sentences, the understanding of melodies requires knowledge that the listener acquires over a long period of time. This knowledge incorporates what the listener knows of the structural frameworks for melodic patterns that are found in the musical culture.
The following three chapters by Dowling, Cuddy, and Bartlett focus on the tonal framework for pitch material in melodies. This tonal framework consists of the do, re, mi pitch pattern that underlies the great majority of European folk and nursery tunes, and that can be found in a set of white keys on the piano keyboard. But the tonal framework is much more than a static set of pitches and the intervals between them, as the chapters indicate. The tonal framework defines certain pitches as stableâpitches where a melody can come to restâand other pitches as relatively unstable and requiring resolution to more stable ones. An example that pervades these chapters is that of a melody perceived as incomplete because it ends on an unstable pitch. Such a melody leaves the listener hanging, feeling a need for resolution.
The example of the incomplete melody illustrates two things about the tonal framework. It makes clear that the framework establishes unstable and stable pitches, with dynamic tendencies attracting the one to the other. It also makes clear that the listener has knowledge of the framework, based on a lifetime of listening to melodies structured in similar ways. That knowledge is largely implicit. In hearing an incomplete melody that ends on re instead of do, for example, average listeners do not say to themselves, âThat melody ended on re and is therefore incomplete.â Rather they experience the incompleteness directly and feel immediately the requirement for resolution. Even the listener with considerable training in music, who can subsequently state the reason for the incompleteness, will experience that incompleteness first before the reason occurs explicitly in consciousness. That immediate experience of incompleteness results from the implicit knowledge the listener has of the tonal framework.
The requirement of a large body of implicit (vs. explicit) knowledge indicates another way in which perceiving a melody is like perceiving a sentence. We have much explicit knowledge concerning the words in language. However, when we hear someone speaking in a language that we know, we do not hear a stream of sounds and then attempt to decipher it in terms of our explicit knowledge. Rather, we hear meaningful words and sentences, already interpreted for us via our implicit knowledge base. Using explicit knowledge is slow and clumsy by comparison, as is immediately apparent in those few cases where we are forced to make use of it; for example, communicating in a language that we do not speak well.
The implicit knowledge that we have of the tonal framework is very slowly built up via perceptual learning throughout our lifetime. The chapters provide evidence that mere exposure to music organized in a particular wayâWestern tonal music for these listenersâis sufficient to induce the auditory system to prepare itself to receive further input conforming to the patterns already experienced. The chapters also provide evidence concerning elaborations of those basic schemes that come about through specialized training in music. The more musical experience the listener has, the more extensive will be the detailed implicit knowledge that listener has of the tonal framework.
The aesthetic implications of implicit knowledge schemes such as the tonal framework are worth noting. In a musical context the tonal scheme defines which pitches are likely to occur and which are not (as Cuddy and Bartlett emphasize), as well as temporal patterns moving from the stable to the unstable and back again. Introducing a pitch foreign to the key defined by the tonal scheme introduces a point of instability that requires resolution, either by a return to the original key or by shifting the tonal center in the direction indicated by the originally foreign note. The aesthetic experience of the listener depends on the stimulus pattern that is heard, setting up a context and then introducing a foreign element and then resolving the situation. But it also depends on the knowledge of such situations that the listener has derived from countless similar experiences, which defines for the listener what will be heard as a foreign or unstable pitch, and what will count as a resolution. There is a delicate balance here between the rule-governed expected pattern and its violation. Without expectancies based on previously encountered regularities the listener will not register the presence of instability, and if too many instabilities are introduced it will begin to break down the regularity of the stable patterns. On the other hand, if no instability ever appearsâif everything always follows the rulesâthe music will be dull and uninteresting.
1
Procedural and Declarative Knowledge in Music Cognition and Education
W. Jay Dowling
University of Texas at Dallas
The term music cognition can refer to a wide variety of processes that go on when a person listens to music, and so at the outset I place some of those processes in perspectiveâplace them in relation to each other and in relation to the broader context of human cognition. To do this I start with a basic situation in which someone is listening intently to a piece of music, without distractions and without any particular aim except to listen. Even this situation is psychologically complicated, but there are several aspects of it that can be described in a general way. Our listener is very likely consciously attending to the patterns of sound at some structural level: either to the broad, global sweep of the sound textures, or to local details. (It would be hard for us to know, however, to what the listener was attending.) The listener is also likely to be consciously or subconsciously developing expectancies concerning upcoming events in the piece. (We can tell that the listener is doing this if there are occasional surprises in the course of listening.) It also seems likely that the listener will be understanding certain aspects of the musical patterns as the piece progresses, in the sense of noticing consciously (or subconsciously taking account of) points of closure, contrasts among the musical elements, and relations among melodic and rhythmic subpatterns. It might also happen that, following the piece intently and developing expectancies, our listener would begin to generate emotional responses to the music. At least, this is what theories such as that of Meyer (1956) suggest, as well as contemporary psychological theories of emotion (Dowling & Harwood, 1986). (In fact, we might imagine in the absence of such responses that the listener might not have been listening so intently.)
This list of aspects of cognition involved in a basic listening situation suggests a multiplicity of cognitive processes occurring simultaneously. Sometimes those processes are conscious, but at other times they are not. For example, sometimes the expectancies that develop will be conscious, and the listener will be consciously anticipating what will happen next. At other times the listener will not be consciously aware of the expectancies, but will only become aware of having had them as a result of little surprises that occur. This can happen, for example, if a piece fails to end in the expected way (see Fig. 1.1).
Another way in which the cognitive processes differ among themselves is that some of those processes depend more on the knowledge the listener brings to the situation, and other processes depend more on what happens in the musical pattern itself. Clearly the process that develops expectancies must rely on both the listenerâs knowledgeâconscious and explicit knowledge as well as subconscious and implicit knowledgeâand the listenerâs perception of the musical patterns in order to project present patterns into future possibilities.
PROCEDURAL AND DECLARATIVE KNOWLEDGE
In this chapter I focus on the role of the listenerâs knowledge in music cognition. In general, knowledge concerning a particular perceptual domain affects the cognition of things in that domain: aiding in the generation of expectancies, facilitating the perception of expected events, and facilitating memory for events that fit the cognitive frameworks of the domain. For example, consider the role of knowledge in solving a perceptual puzzle such as the âfind six lions in the jungleâ type of picture in childrenâs game books. To find the lions you need to (a) know what a line drawing of a lion looks like, (b) know that there are lions to be found, and (c) have some practice discerning lion drawings in confusing contexts.
We can point out two kinds of knowledge that are involved in perceptual tasks such as finding hidden visual figures and listening to parts intertwined in a complex piece of music. One kind of knowledge is called by psychologists declarative. Declarative knowledge is the consciously accessible kind that you can talk about. You can tell someone: âThere are six lions to be found here.â You can mentally review a list of lion features: large, cat-like creature with furry mane, tawny color, long tail with tassel, and so on. That is all declarative knowledge. But even when you know all this you will still need practice finding the lions before you become rapid and accurate in doing it. In the course of solving a number of problems of this sort you will develop a kind of practical know-how that we call procedural knowledge. Procedural knowledge is often consciously accessible only through its results, and not often explicitly present to consciousness as such.
FIG. 1.1. âTwinkle, Twinkle, Little Starâ with a normal ending (a) and a surprise ending (b).
The contributions of declarative and procedural knowledge to perceptual experience can be seen clearly in listenerâs understanding of spoken language. Knowing in the declarative sense what someone is talking about helps us understand what is being said. But understanding someone speaking requires vast amounts of practice in listening. We have only to try to understand someone speaking an unfamiliar language to appreciate the importance of practice and the procedural knowledge it develops. With a couple of years of academic study of foreign language we can use declarative knowledge to figure out written messagesâwe can look up unknown words and decipher the syntax in an abstract and formal way. But we will continually have to ask people to speak more slowly, because our procedures for automatic perceptual processing of the speech sounds, so proficient in our own language, are not well-developed. This example points out two important differences between declarative and procedural knowledge. Declarative knowledge is easily and rapidly acquired. (How long does it take to learn that George Washington and Josef Haydn were born in the same year?) But declarative knowledge is relatively slow and clumsy to apply. In contrast, procedural knowledge takes enormous amounts of practice to acquire (as in the example of a foreign language), but is rapid and often automatic in application. Automatic here means that it is brought into play in cognitive information processing without conscious intent and without distracting us from attending to other things, including the results of its application.
The roles of declarative and procedural knowledge in cognition have been the focus of a paradigm shift in psychology over the past 10 years or so. When I was in graduate school in the 1960s, models of cognition were heavily based on declarative knowledge. A model of the process of understanding words in a sentence typically contained a module for interpreting speech sounds, turning them into something like a string of representations of letters. Then those strings of sounds would be looked up in a sort of dictionary module and be given a meaning. The meaning they were given would be roughly the same as the meaning the listener would give you in answer to the question âWhat does X mean?â The process of understanding a spoken word would be conceived of as essentially a two-stage process: procedural interpretation of sounds, followed by accessing declarative meanings of those sounds.
Such a model depends critically on declarative knowledge. Declarative knowledge forms the basis of much of our system of education, both ideologically and economically. We live in a logocentric culture, to use Clarkeâs (1989) term, and we rely heavily on the transmission of information organized in words and symbols embedded in syntactic systems. Our whole method in training the young as well as training the not-so-young is based on the efficient transmission of declarative information to large groups of people at once. The culture is well characterized as logocentric because we tend to think we only really understand something when we can explain it verbally, that is, in an explicit declarative way. Such declarative knowledge can contribute to expectancies. Sometimes the declarative component parallels the procedural, as in the case where we know that at the end of a piece in European tonal music a dominant harmony typically resolves to the tonic, and the violation of that expectation in Fig. 1.1 strikes us on both levels. And sometimes declarative knowledge operates more independently of procedural, as when we know that most classical concertos end with a rondo, and thus expect one at the end of a concerto we have never before encountered.
We contrast this declarative approach with a model of cognition based on a procedural approach. In this model, the brain, faced with patterns of stimulation, finds effective ways to process them. It takes advantage of regularities and redundancies in the patterns. If some pattern...