Several of my undergraduates every year gleefully sing to me the “Parts of the Brain” as they learned them from this show, set to the melody of “De Camptown Races,” with a high-pitched voice chiming in the words “brain stem” where the “doo-dahs” would go: “neocortex, frontal lobe (brain stem, brain stem)/hippocampus, neural node, right hemisphere.” By combining laboratory experiments with real-life ethnographic studies, psychologists are just beginning to understand how these and other parts of the brain are able to encode and preserve so much information in music.
In the 1930s, Albert Lord and Milman Parry recorded folk songs in the mountains of (then) Yugoslavia (but not Crete, Mauritania, then Transylvania). For hundreds of years or longer, traveling singers there, mostly Muslim, have gone from town to town, staying for a few days at a time and singing oral histories that can be thousands of lines long and take several evenings to complete. Experienced singers in this tradition might know thirty to a hundred such epics. Some of them memorize their songs with very high accuracy: In one case, a singer heard a song once and then brought it into his repertoire; tape recordings seventeen years later show astonishing consistency, with only very minor errors of wording.
The Gola of West Africa place a particularly high value on the preservation and transmission of tribal history. Part of the reason for maintaining the histories is practical. The knowledge of kinship origins can help establish among contemporaries familial connections and the reciprocal responsibilities attendant to those; being able to claim a relative when you have no food can mean the difference between life and death in a subsistence culture. In addition, ancestors are viewed as distinct personalities who continue to exert influence over the course of present events, even when they’ve been long dead. Much of this rich oral history is held in musical form by oral Golan historians.
The ancient Hebrews set the entire Torah—the first five books of the Old Testament—to melody and recalled it from memory for more than a thousand years before they ever wrote it down. Even today many Orthodox rabbis can sing every word by heart. The so-called Oral Torah, the commentaries, instructions, emendations, and explanations contained in the written Talmud, have also been memorized verbatim by many, also set to music.
Songwriters know implicitly that setting something to music is the best guarantee that it will be remembered. In contemporary society, writing things down on paper, a PDA, or a computer may seem more practical, but it may not be more powerful. Songs stick in our heads, play back in our dreams, pop into our consciousness at unexpected times. Oliver Sacks tells the story of a song he could not get
unstuck
from his head, Mahler’s
Kindertotenlieder
(songs on the death of children). Unable to identify the piece—which grew to create in him a sense of “melancholy horror”—Oliver sang it for a friend. “Have you abandoned some of your young patients, or destroyed some of your literary children?” the friend asked. “Both,” Oliver answered. “Yesterday I resigned from the children’s unit at the hospital . . . and I burned a book of essays.” Oliver’s mind had brought up Mahler’s song of mourning for the death of children, a reflexive, inventive, and subversive way of symbolizing the previous day’s events.
When Genesis sings “All this time, I still remember everything you said,” or when Bryan Adams sings “I remember the smell of your skin/I remember everything/I remember all of your moves,” they are encoding their personal memories in song. By bonding the sentiments to evocative chords, rhythms, and melodies, they are tapping into a millenia-old and powerful way to save primitive, pure emotional reactions to feelings and events that are important to them. The music itself, even without words, can be very effective at evoking the right sentiment. (Research shows that people are very accurate at identifying the intended emotion of a piece of novel instrumental music.) Adding in the words and their interaction with melody, harmony, and rhythm can print the message indelibly in memory for a lifetime; even longer as new generations of listeners hear the song, learn it, and pass it on.
Some knowledge songs are written as a plea for the listener to remember something specific and important to the writer, as in “The Last Song,” by indie rockers Sleater-Kinney:
I need you out of me before I turn into you
I can’t stand to look at you,
Until you remember everything
The recipient’s only hope for saving the relationship is to “remember everything,” and the song serves as a reminder that this is what (s)he needs to do.
Often songwriters are not asking for others to remember, but write
for themselves
to remember. Take one of Johnny Cash’s most famous songs, “I Walk the Line”:
I find it very, very easy to be true
I find myself alone when each day is through
Yes, I’ll admit that I’m a fool for you
Because you’re mine, I walk the line
On the surface, it seems like a sweet love song, sung by the man to his sweetheart back home. But in the third verse, Cash explicitly invokes memory as he pledges to think of her:
As sure as night is dark and day is light
I keep you on my mind both day and night
And happiness I’ve known proves that it’s right
Because you’re mine, I walk the line
My reading of this verse is that Cash is not in fact singing the song to
her,
but to
himself
. The ironic underpinning is that in fact he
doesn’t
find it very, very easy to be true. He
wants
to be, but it is a struggle. The song’s function is to remind
him
why he’s doing this—because of the “happiness [he’s] known” with her and that he doesn’t want to risk. “I Walk the Line” is the song of a man in conflict, on the road, with a wandering eye, hoping against all odds and knowledge of his own weakness that he will be true. “I Walk the Line” is a knowledge song, a song of feelings encoded in song so that he will remember “when each day is through” what he promised himself he would do.
Popular songs such as “I Walk the Line” or Genesis’s “In Too Deep” are only a dozen or two dozen lines long. How are Homerian epics, or the long oral histories and ballads of the Yugoslavians, the Gola, or the ancient Hebrews, remembered? Psychologists Wanda Wallace and David Rubin, among others, believe that the mutually reinforcing, multiple constraints of songs are crucially what keeps oral traditions stable over time. In most cases, it turns out, the songs are not remembered verbatim, word for word. Rather, broad outlines of the story are remembered, perhaps using visual imagery, and structural constraints of the song are memorized. This is a much more efficient use of memory than pure rote memorization of the words, using up far fewer mental resources. In Chapter 1 I spoke about the importance of form in poetry, and in song, form is
the
critical feature that helps to recall lyrics.
The mutually reinforcing, multiple constraints that help us to remember song lyrics are principally rhyme, rhythm, accent structure, melody, and clichés, along with various poetic devices such as those we saw in Chapter 1, including alliteration and metaphor.
The rhyming scheme we find in most songs constrains the words that can appear in the last position of rhyming lines. Even though there may exist several words rhyming with the correct word, semantic constraints will prevent most of those words from working in the context of the song. Say, for example, I know a song that begins:
The sun goes down, old friends drink and chat
The wind’s in the trees, the dog growls at the——.
Although words such as “hat,” “fat,” and “mat” could fill in the blank and maintain the rhyme, our brains reject them almost instantly, unconsciously, because they are considerably less
likely
than an alternative (“cat”) in terms of semantics, not to mention in terms of our life experience of storytelling norms in our culture, in which cats are far more frequently mentioned in the same sentence as dogs than are hats, chats, mats, gnats, wine-making vats, or baseball bats. All work syntactically and poetically; one is far more likely semantically.
In Rodney Crowell’s song “Shame On the Moon,” suppose you just can’t remember the word at the end of the first line in the second verse, and all you can remember is that the lines rhyme with each other (a structural memory):
Once inside a woman’s heart, a man must keep his——
Heaven opens up the door, where angels fear to tread
There are perhaps a dozen or so words that rhyme with tread: bed, bread, red, shed, but none of these seem likely in terms of semantics. Even if you can’t remember the correct word (“head”), your brain can figure it out more or less on the spot during reconstruction of the song. In fact, there now exists a wealth of research that what we think of as our ability to remember is grossly over-inflated, and that many dozens or hundreds of times a day we are
creating
bits and pieces of recollections on the spot. Our brains stitch these creations together seamlessly with our actual recollections and we’re none the wiser. In the jargon of cognitive science, this is called the constructive aspect of memory, and it happens so often, so spontaneously, and so quickly that we usually don’t know which parts of a recollection are faithful copies of our memory and which are simply plausible inferences our brains made for us. If bits and pieces of a song lyric can be generated on the spot, according to rules, then less of the song actually has to be remembered, and this is far more efficient for the brain.
We engage in this kind of parsimonious and rule-based memory retrieval all the time in other domains, for example in memorizing phone numbers. I have several friends in San Francisco and I’ve memorized their seven-digit phone numbers. When I go to call them, I remember that they’re in San Francisco, and that the area code for San Francisco is 415. Instead of having to memorize an extra three digits for all of my friends, I memorized the
rule
(San Francisco: add 415 to the number). I didn’t do this consciously, this isn’t some sort of mnemonic
strategy
I employed—it is automatic.
An example of a constructed memory is if I were to ask you about the last time you went to a restaurant. You might respond that you walked in the door, were met by a host or hostess, shown to your seat, where you were given a menu, your order was taken, your food was brought, and so on. Now suppose your narrative doesn’t mention the server bringing you the bill at the end of the meal, and I ask you, “Did the server bring you the bill after you were finished eating?” In laboratory experiments like this, many people say yes, they remember being brought the bill, but there exists ample evidence that they aren’t actually
remembering,
they’re merely
assuming
(or “constructing” the memory in the jargon of cognitive science), because this is part of our shared, common knowledge about what types of things usually happen in restaurants. We don’t have to remember all the events for any specific restaurant visit, because some of them are so similar and tightly scripted (and indeed, we tend to remember only if something distinctive occurs, such as the server bringing you the
wrong
bill).
Interestingly, we engage in a very similar process when we recall a conversation, piece of text, or speech, using what psychologists call
gist
memory. We tend to recall the gist, but rarely the precise words. Again, this is a demonstration of parsimony in memory and the fact that in a world of constantly changing environments, literal recall is seldom important. Most messages then are encoded using only a few words or concepts, and our knowledge of the English language and how to form sentences allows us to recreate something
similar
to what was said—close in meaning, but perhaps different in specific expression. We also do something analagous with melodies. When I was in a high school marching band, we played John Philip Sousa’s famous march, “The Stars and Stripes Forever.” There are several points in the piece where the wind instruments play a low note, followed by a rapid flurry of notes all leading up to a much higher note. It would be inefficient and also unnecessary to try to memorize every single note of that flurry. Instead, research shows that in cases like this, musicians typically memorize the low note, the high note, and how many beats there are available to get from one to the other. Then, using their knowledge of scales and tonality—
rules
—they construct the intermediate notes as and when they’re needed.
The very poor typical recall of text stands in stark contrast to the very good typical recall of song lyrics, especially in the case of long epic ballads and the information-set-to-music that I call knowledge songs. Again, this is because songs provide form and structure that jointly serve to fix and constrain possible alternatives. We don’t need to store every word of the lyric in our brain’s memory banks; we only need to store
some
of them, along with the story and knowledge of the structure of the song. Structural knowledge may include things like the rhyming pattern (the fact that lines one, two, and three all rhyme while four doesn’t; or the fact that line one rhymes with line three and line two rhymes with line four, and so on).