Wednesday, December 17, 2008

Music Can! Music Can't!

Memory for Music: Effect of Melody on Recall of Text
Wanda T. Wallace
Journal of Experimental Psychology: Learning, Memory, and Cognition 1994, Vol. 20, No. 6, 1471-1485

Lexical Learning in Sung and Spoken Story Script Contexts
Theresa A. Kouri and Jennifer Winn
Child Language Teaching and Therapy 22, 3 (2006); pp. 293-313

Review and Response by John Picone

Music and language, it would appear, are inextricably entwined in the development of the child. There are so many “linguistic” elements of music that influence linguistic development. Musical notes are like phonemes; musical pitch and rhythm are like intonation; melody contours are like syllables in words. There are many studies that show how powerfully music influences our phonemic awareness, our sensitivity to emotion and attitude expressed in intonation, our ability to segment spoken words.

These two studies also look at music’s influence on different but related aspects of language learning. In the first study, Wallace hypothesizes “that the melody will facilitate learning and recall of the text above and beyond what is contributed by the rhythm and rhyme present in the text” (Wallace, p. 1472).

For Wallace, “the underlying notion is that the melody provides rich information about the features of the text as well as a direct connection between components of the melody and components of the text. These connections then are access points or cues to memory. Thus, thinking about some component of the melody will cue the parallel component of the text” (Wallace, p. 1472).

Wallace conducted four experiments with the participants asked to recall the text of a ballad containing 80 – 85 words in three verses. The participants, in two randomly selected groups, heard sung and spoken versions of the ballad.

In the first experiment, the participants heard the ballad five times and were asked to recall the text after the first, second and fifth repetitions, then, in a delayed-recall task after about 20 minutes. There was markedly better recall of the sung version.

The second experiment sought to isolate the melody as a recall component. Half the participants were exposed to the sung version of the ballad as in the first experiment. The other half heard the ballad “spoken with a rhythmical intonation, emphasizing the rhythmically stressed syllables. In the background, a metronome tapped in synchrony with the verses” (Wallace, p. 1476). Again, though not as great as in the first experiment, recall was significantly better with the sung version.

In the first experiment, the participants had ample opportunity to learn the melody as there were three verses. That is, the participants heard the melody three times as often as the text. In the third experiment, only one verse was heard by the participants; the melody was heard as often as the text. Wallace hypothesized that “if the melody was not so well learned as in Experiment 1, the perhaps it would not serve as an adequate encoding or retrieval cue and might actually serve as a distraction” (Wallace, p. 1477). This, indeed, proved to be the case. Recall of the spoken text was significantly greater than the one-verse sung version.

The purpose of the fourth experiment was to replicate the findings of experiments 1 and 3 and establish further evidence that music contributes more than rhythm in its ability to facilitate recall of text. In this experiment, there were three groups of participants: the first group was exposed to the ballad as in the first experiment; the second hear the ballad in a spoken version; the third group heard the three verses of the ballad with three different melodies. The recall the text with the repeated melody was significantly higher than the other two versions. There was no significant difference between recall of the spoken text and the three-verse text presented with three different melodies.

Wallace’s conclusion: “The melody of a song can indeed make a text more memorable as compared with hearing the text out of the context of the melody, at least as long as the melody is simple and easy to learn” (Wallace, p. 1481).

The Kouri and Winn study also explores how music might assist in language learning. Indeed, the study refers to the research done by Wallace in the study noted above. Kouri and Winn do not explore music’s influence on the recall of text, but on the learning of new text.

The purpose of this study was to provide evidence regarding the effects of musical input on the quick incidental word learning skills (QUIL) of preschoolers with developmental and language delays. Children were presented with novel vocabulary items in two scripted stories that were presented via sung and spoken input while observing brief story enactments. The first goal of the study was to establish whether young pre-schoolers with mild DD (developmental delay) and SLI (specific language impairment) are able to acquire novel lexical terms after limited exposure to short, sung and spoken story scripts. Then, it was determined if children’s comprehension or production of novel lexical terms varied as a function of exposure to sung versus spoken script presentations over two experimental sessions. It was hypothesized that children in this study would demonstrate increase QUIL in the sing script context, given previous research outcomes regarding the use of music and word learning. (Kouri & Winn, p. 297).

The participants in the study were 16 children between the ages of 3.6 and 5.1 years of age, with an average age of 4.1. They all demonstrated expressive and/or receptive language delays of at least 12 months.

Drawing upon Wallace’s conclusions about the role of a repeated, simple melody, Kouri and Winn set their scripts to the tune of “Down By The Bay,” a well-known children’s song popularized by Raffi. The melody was repeated for each of the three verses of two created story scripts, “The Lake” and “The Playground.” Eight novel lexical items were created for use in the scripts: two syllable nonsense words comprised of early occurring phonemes and syllable types. The four novel words for each story were embedded three times each in utterance positions of emphasis such as at the beginning or end of sentence lines. Novel target objects were created for each neologism and manipulated as the story was acted out with toy characters and other objects, such as furniture, against a hand-sketched backdrop of the setting. Each participant heard each story in both sung and spoken conditions. For example, a child would, in the first session, hear “The Lake” spoken and then “The Playground” sung, and in the second session, this would be reversed.

Child participants were instructed to listen to the story and watch the experimenter as she acted it out three different times, twice with the live voice presentation and once with the audio-recorded version of either the sing or spoken script, depending on what condition was being presented. Then children were allowed to manipulate story characters and objects while listening to two more audio-recorded presentations of the sung or spoken script. If a child diverted attention away from the story context, the experimenter would point to target objects as they were presented in the story line. If this did not regain the child’s attention, the experimenter stopped the story and resumed again after refocusing the child (Kouri & Winn, p. 300).

After a five-minute break, three lexical probes were administered: production, comprehension and generalization. During the production probe, each of the four novel objects were presented to the child who was asked, “Tell me what this is called,” or, “Do you remember the name of this one?” In the comprehension probe, the child was presented with the novel object along with three distracters including another novel object from the story. The participant was instructed to “find the ____.” The generalization probe was similar to the comprehension probe except that each novel object was presented with three distracter foils, objects that closely resembled the novel object, differing in size, shape or colour. Again, participants were asked to “find the _____.”

It should be noted here that there were many instances of unsolicited production of the novel words by the participants. The researchers, however, could not be certain whether these were being used as labels for the novel objects or merely as imitation.

By way of conclusions, this latter phenomenon suggested to the researchers that new lexical items, after limited exposure in both sung and spoken texts, could be acquired by children with SLI and DD. However, “quick incidental learning was not differentially affected in either context. In terms of lexical comprehension, it appears that the melodic input accompanying script enactments failed to facilitate increased recognition or receptive identification of target and generalization lexical items” (Kouri & Winn, p. 304).

In presenting implications for practice, the researchers, with clear reference to some kinds of music therapy interventions, indicate that “a practitioner should not assume that children with language and mild developmental delays will tend to comprehend or name (upon demand) more new lexical items just because they are presented in a sung context” (Kouri & Winn, p. 307).


My first response to the Kouri and Winn study is something of admiration for the integrity of researchers. Without hesitation or qualification, they concluded that there is something that music cannot do: assist in learning new lexical items – words! My admiration, though, is qualified in that I believe there may be some significant flaws in the design of their experiment. Such an experiment, it seems, would need the prerequisite of a sound understanding of linguistics, specifically language acquisition. It is questionable whether the researchers brought such knowledge to this experiment. Did they consider, for example, how a child learns any new word? What is the context that facilitates this phenomenon?

The researchers truly extended themselves in enhancing the opportunity for the participants to learn the new lexical items: repetition, enactment, redirecting attention. In fact, one wonders to what extent these elements of the experiment distracted from the learning rather than assisting it. However, an essential element to learning a new word – or genuinely learning anything, for that matter – is not present in this experiment: need. Consider what one might reasonably assume to be the first lexical items any normal child learns: “mommy,” “no,” “mine.” Such words are useful tools to the child; learning them is important.

The context in which these children are potentially learning new words is artificial, much the same as grammar drills, vocabulary lists and phonics exercises. There may certainly have been a “fun” factor which may have enhanced learning, but this is not accounted for by the study.
The Wallace experiment is likewise artificial. The real context of recalling the text of a ballad would take us back to a pre-literate society when passing along the stories and legends of a society through song was part of the culture and, as few could read or write, this was the only way to preserve such stories. But in the Wallace experiment, where music is used to facilitate the recall of text rather than the learning of new text, is done with adults. There is a deliberateness to their recall of which, perhaps, only an adult is capable. These same two experiments may have yielded interesting results had the two groups of participants been reversed.

Kouri and Winn make several references to the Wallace study and, it would appear, the findings of the latter lend integrity to the hypothesis of the former. However, even though Wallace states in her hypothesis “that the melody will facilitate learning (emphasis mine) and recall of the text” (Wallace, p. 1472), there is no actual “learning” in this experiment. In the three ballads used in the experiment, there is only one word that might be foreign to the participants: “knocker,” as referring to the brass ring tapped against a door to announce one’s presence. How might the Wallace study have turned out had there indeed been some new lexical items to “learn”?

Another question I have about the design of this experiment is what seems to me to be an inherent contradiction in the QUIL method, especially with children with DD and SLI. I agree that all genuine learning is, indeed, incidental: we’re not aware that we’re learning at all! The problem here, it seems, is in the concept of “quick.” Aside from the issue of meaningful context, how many times does a child need to hear a word to “learn” it? Growing up in an Italian home, “pasta” and “sugo” (my mother’s word for spaghetti sauce) were words I heard often and in a real, “whole language” context. But how often did I hear them before they became “learned”? This leads me to question whether or not the children in the Kouri and Winn study who, through the lexical probes, gave correct responses, have genuinely learned the new lexical terms. What if they were asked the same questions a week later? One would speculate they would have completely forgotten such terms as they were unlikely to hear them again outside the experiment situation. So, what, in fact, constitutes “learning” a new word?

There is much research that strong suggests that, if the mechanisms for learning language and music are not the same, they are at the very least significantly intertwined. I don’t feel that the conclusions of this study are credible enough to suggest that there is a distinct neuromechanism for lexical comprehension.

As always, there’s much more work to be done.

PS – It was through “Down By the Bay” that my own children learned the word “watermelons.” Four syllables! But then, it was summer and we had it for dessert a lot and were growing some in the garden.

1 comment:

Lee Bartel said...

This is an excellent overview of this set of studies. I agree with you on the lack of understanding of language learning. You cannot insert what amounts to meaningless vocables - gibberish to the child - and expect it to be learned. We learn words as meaning. That is the central flaw of phonetics only reading approaches. Kids do not learn language as disconnected sound and do not learn to read with the assembly of meaningless sounds as phonemes. But the learning of melodies can in some ways be similar - melodies that are syntactically sensible within a musical culture (language)are learned more easily that what amounts to a random set of pitches. Hence the difficulty of people learning to understand and therefore like much of the music of the early part of the 20th century.

Thanks for this John