The reality of these chunks has been demonstrated many times in psychology laboratories. Asked to sing song lyrics from memory, beginning from an arbitrary point, people have great difficulty. Asked even to answer simple questions about lyrics they know, people are influenced by the hierarchical structure of the lyrics—revealing in the laboratory certain organizational properties of human memory. Here’s an example. Does the word my appear anywhere in the lyrics to the song “Hotel California”? What about the word welcome? Both words do appear, my is the ninth word in the song, and welcome is the ninety-sixth word. But people take longer to say yes to the first question than to the second, and psychologists believe this is because welcome is the first word of the chorus, a privileged position in the hierarchy of your memory for the piece.
Most North American children learn the alphabet by learning the letters set to the melody of “Twinkle Twinkle Little Star” (the same melody as the beginning of “Ba Ba Black Sheep”). The song has phrase boundaries because of its rhythmic structure, gaps between the letters g and h, k and l, p and q, s and t, and v and w, forming natural “chunks”:abcd efg hijk lmnop qrs tuv wxyz
Indeed, most children don’t memorize this all at one sitting, but rather they work their way up, memorizing these small units. The rhyming scheme helps too: The ends of all the chunks rhyme with each other except chunks three and five. Even though as adults we know the letters of the alphabet (or think we do), many of us still rely on that song when searching for the specific location of a letter. In one experiment, it was found that it takes college undergraduates much longer to say what letter comes just before h, l, q, or w than before g, k, p, and v. Crossing the chunked boundary carries with it some cognitive cost.
Notwithstanding what I said about errors at phrase boundaries, some professional musicians and Shakespearean actors do indeed have perfect recall for a memorized string and can begin anywhere; this ability probably develops from overlearning processes, and under stress it may disappear. Something like it shows up in even semiprofessional or amateur actors and musicians when they need to start from the very beginning of a piece. Effectively, their “stitching together” of subcomponents has worked so well that the entire piece, though at one time consisting of small units, has become a single memory trace, not easily disassembled, occasionally indestructible, lasting a lifetime, and sometimes outlasting even the memory of their own family members’ names.
The use of chunking and the occasional inviolability of long, overlearned sequences is not confined to contemporary Western society. The Greeks discussed these ideas in formulating instructions for mnemonics two thousand years ago, and the anthropologist Bruce Kapferer (from the University of Bergen in Norway) has observed them in researching the myths of Sri Lanka. In trying to form a catalog of different demons and their characteristics in a cultural group he was studying, he would ask a local oral historian to describe the myth of a particular demon. The historian would respond that the exact details were contained in a certain song. “I will sing it and you tell me when the demon you want has his name mentioned. Then I will go slow so that you can put it onto your tape recorder.” The myth information was stored in the song, and the song was known only sequentially, and only from the beginning.
Getting back to the Greeks, the 2,500-year-old Iliad and the Odyssey represent great feats of memorization, without music, but with clear poetic, rhythmic constraints doing much of the work and thus creating less demand on the brain. Their prosody is very tightly constrained. As just one example, the number of syllables per line is almost always constant, and the last five syllables in a line are almost always long-short-short followed by long-short. The ordering of short and long syllables, and the preferred locations for word breaks are formulaic—not just any word will fit the rules. For example, words containing a long-short-long or short-short-short syllabic structure can’t be used in Homeric epic at all. Obviously, if one has learned the rules for such a form, opportunities for inserting the wrong words are extremely limited.
According to Jewish tradition, the complete Torah (the first five books of the Old Testament) was completely memorized by Moses, and then taught to the elders and leaders of the Hebrew people in the Sinai Desert, who in turn taught it to the million or so people who left Egypt as part of the Exodus, sometime around 1500 B.C.E. We know that the Hebrews had written language (the tablets of the Ten Commandments were written), but on Moses’s strict instruction, not one word of the Torah was to be written down, and for more than one thousand years the history, knowledge, religious customs, and practices were reportedly handed down only through oral transmission. And the form of that oral transmission, according to all accounts, was song.
Jewish mystics believe that the very sound of the words will bring divine favor, even if the speaker doesn’t understand the words themselves. Similarly, in the Zoroastrian tradition, it is believed that the soul (Urvaan) can be reached by the specific vibrations that come from chanting the Avesta Manthras. It’s not just the meaning of the prayers, but also their sound that matters, for the “attunement” of the soul. In Zoroastrianism, Staota Yasna is the theory of auditory vibration. That’s why the prayers are still recited in the original language, Avesta, even though it isn’t spoken any more.
The Qur’an was also set to rhythm and melody—a chant—and learned in a similar fashion, although it is explicitly not considered “music” and is structurally very different from Arab music we hear today; in fact singing the Qur’an is strictly forbidden. The Qur’an itself describes the means of its recitation (tarteel) in verse four: “and recite the Qur’an in slow measured rhythmic tones.” The power of song to aid memory is evidenced in a following fatwa against singing. Islamic scholars believe that “music and singing carrying obscene content, instigating to sin, lechery, destroying noble intentions and leading into temptation, are inadmissible (haram). . . . The degree of inadmissibility becomes higher if an obscene vocabulary acquires a musical accompaniment that contributes to a better remembering and thus enhances its impact.”
In the case of the Torah, the melody itself contained clues not only to the words, form, and structure of the narrative, but also to interpretation of words or passages that might otherwise be ambiguous. That is, the assignment of words to melody (and vice versa) was not arbitrary—it helped not only as an aid to memorization and recall but also to ensure the correct interpretation.
This sort of interplay is not at all unusual 3,500 years later. In the song “Superstar” as sung by the Carpenters (written by Leon Russell), Karen Carpenter sings the line “Long ago, and oh so far away” using a vocal technique that artfully reinforces the meaning of the words. She delays the pronunciation of the word “far,” reinforcing the idea of distance. While holding the word “away,” she brings out a subtone in her voice that conveys the sense of deep loss and separation. In Steve Earle’s “Valentine’s Day,” the singer is surprised by the arrival of the day, and realizes too late that he has forgotten to get his girlfriend a present. He writes her a song instead. The appearance of surprising and nonstandard chords underscores the meaning of the words, adding tension and deeper meaning to the lyric.
The reasons for the insistence that the Torah be transmitted orally are a matter of speculation. One proposal is that the success of the Jewish people, one of the oldest continuously living civilizations in the world, is due to especially close ties between parents and children and the bond created through the oral transmission of knowledge—knowledge which, in the case of the Torah, intimately binds family history, moral lessons, political history, codes of daily conduct, and instructions for maintaining an orderly, just society. If the information had been written down and learned by reading, the knowledge transmission would have flowed in one direction, from book to student. The oral transmission enabled—virtually required—interaction, questioning, participation; what the physicist-turned-Torah-scholar Aryeh Kaplan called a living teaching. Indeed, the ancient Hebrew scholars wrote that “the T
orah is meant to be alive, to be spoken.” Like the poetry we encountered in Chapter 1, it is meant to be heard, both in the ears of the people and in the minds of those who have learned it and can play the song of it back in their heads at will, in times of scholarship, trouble, or praise. It’s also been proposed (somewhat less logically) that the restriction on writing the Torah down existed because some knowledge of customs and traditions would be lost if it was committed to writing—that the sum total of the knowledge known by the people exceeded what could be written down. When the rabbis decided sometime between 150 B.C.E. and 200 C.E to commit all the teachings to writing, much debate and disagreement indeed ensued, about many details. (All of the debate is captured in the Talmud—in fact, that is primarily what the Talmud is—a record of what were essentially judicial proceedings and deliberations about what precisely the oral teachings were and how they were to be interpreted.)
From a memory standpoint, the cantillation (as the Torah melody is known) provides the same sorts of constraints that other songs do—perhaps even more—facilitating the memorization and preservation of an enormous amount of text. Without recordings, however, it is impossible to know for sure how well the original words of sacred texts such as the Torah and the Qur’an were preserved through their oral transmission. We don’t know, and can’t know, the extent to which melodies changed, rhythms were rewritten, emphases were altered. But the fact that different subgroups of contemporary Jews sing different melodies suggests that there was no one magic formula for preserving the information—time and tide would have caused minor changes as in the child’s game of telephone, and over generations these differences could have become considerable, and they would have become amplified. Humans are a highly adaptive species. As we moved to new locations, to communities with their own musical cultures and traditions, original melodies may well have become altered or distorted by the influence of local songs. Even the prosody of the new languages (the “music” of the language) has been shown to influence the songs of that linguistic culture. As the Mongols entering southern Europe on horseback, the Armenians dispersing to Paris, and Italian-Americans crooning in Hoboken found, local sounds pull at the immigrants’ long-preserved melodies and rhythms, creating new hybrids that continue the cultural evolution of their songs, at the possible expense of losing some of their original (melodic and textual) information.
The existence of different melodies for Torah today suggests that errors may indeed have crept into the text when it was transmitted orally—if melodies can change, so can words. (In fact, much of the discussion during the compilation of the Talmud in the first few centuries C.E. acknowledged that some errors by then had already crept in, and concerned how to resolve those errors.) Indeed, the discovery of the Dead Sea Scrolls reveals that multiple versions of the sacred texts exist. From both a cognitive and a theological perspective, the errors are mostly relatively minor and unimportant, of the type we saw in Rubin’s study of ballads.
Across all these examples, a common thread emerges: Knowledge songs tell stories, recount an ordeal, a saga, a particularly noteworthy hunt—something to immortalize. The demonstrated power of song-as-memory-aid has been known to humans for thousands and thousands of years. We write songs to remind ourselves of things (as in Johnny Cash’s “I Walk the Line”) or to remind others of things (as in Jim Croce’s “You Don’t Mess Around with Jim” or Fleetwood Mac’s “Don’t Stop”). We write songs to teach our young, as in alphabet songs and counting songs. We write them to encode lessons that we’ve learned and don’t want to forget, often using metaphor or devices to raise the message up to the level at which art meets science (rather than simple observation), making it at once more memorable and more inspiring, as Andy Partridge does in “Dear Madam Barnum” (performed by his band XTC). The name Madam Barnum is clearly made up, meant to portray her as the ringleader in some abusive emotional circus she has subjected the poor songwriter to, and from which he now hopes to extricate himself:I put on a fake smile
And start the evening show
The public is laughing
I guess by now they know
So climb from your high horse
And pull this freak show down
Dear Madam Barnum
I resign as clown
Songwriters often invoke a well-known tale or legend in the context of a new song. In “Dear John,” the writer of this song (Aubry Gass, and as sung by Hank Williams) is again encoding an experience into a song, presumably so that he won’t forget that it was his own misbehavior that caused his woman to leave him; he weaves into the message two well-known Old Testament references:Well when I woke up this mornin’,
There was a note upon my door,
Said “don’t make me no coffee babe,
’cause I won’t be back no more,”
And that’s all she wrote, Dear John,
I’ve sent your saddle home.
Now Jonah got along in the belly of the whale,
Daniel in the lion’s den,
But I know a guy that didn’t try to get along,
And he won’t get a chance again,
And that’s all she wrote, Dear John,
I’ve fetched your saddle home.
Note the interesting shift to the third person in the second verse (“I know a guy that didn’t try to get along”), a conceit to make us think that it isn’t actually he who was left by a woman, a move that underscores the point that he’s sending the message out to others as a warning: Don’t do what I did; treat your woman right.
Hard-won lessons are a staple of knowledge songs, from Paul Simon’s “Run That Body Down” to Ani DiFranco’s aptly titled “Minerva” (after the Roman goddess of knowledge) to the Magnetic Fields’ “You Love to Fail.” As with Aubry Gass’s, the songs seem to be simultaneously directed from the songwriter to him- or herself and to all of us. Guy Clark, one of my all-time favorite songwriters, brings a lifetime of lessons seemingly learned the hard way into his song “Too Much,” a romp that is made all the more fun (and memorable) by the form he imposed on himself: Every line of the verse begins with the same two words (“too much”), compiling a litany of everyday pleasures, too much of which will cause the various calamities specified at the end of each line:Too much workin’ll make your back ache
Too much trouble’ll bring you a heartbreak
Too much gravy’ll make you fat
Too much rain’ll ruin your hat
Too much coffee’ll race your heart tick
Too much road’ll make you homesick
Too much money’ll make you lazy
Too much whiskey’ll drive you crazy
. . .
Too much limo’ll stretch your budget
Too much diet’ll make you fudge it
. . .
Too much chip’ll bruise your shoulder
Too much birthday’ll make you older
Part of what makes the song memorable is the obvious fun that the composer had in writing it, reflected in the joy the performers bring to playing it. The sense of whimsy is enhanced by decomposing the familiar idiom “carrying a chip on your shoulder” to yield “too much chip’ll bruise your shoulder.” In “too much limo’ll stretch your budget,” he taps into our memory associations of limos and “stretch limos” to use both senses of the word stretch. (It is no wonder that Clark is a favorite of many of the best songwriters in the business, including Rodney Crowell, whom Guy mentored.)
A song like “Too Much” turns the memory process into a game—the first part of each line cuing the second part or vice versa. If we forget the line, logic can often deliver it to us just as it does in “I Know an Old Lady Who Swallowed a Fly.” It is worth making the cultural point that these sorts of songs are found in every society we know of. And the fact that children tend to love these songs is evidence that our ancestors found this sort of mental play both rewarding and an efficient form of learning and information transmission.
Up until this point, I’ve been conside
ring songs as they are recalled and sung by one individual at a time. But knowledge songs—from Huron’s yellow school bus songs to Torah cantillation—are more typically sung by groups of people. In this context their position as a foundation of culture and their durability become even more apparent. I’ve already described the social bonding that comes from synchronous music making, and the neurochemical effects of singing, but there are manifest cognitive benefits that are conferred to the group-as-a-whole, apart from any benefits to the individual when people sing together. Group singing shows a special ability in retrieving information that a lone individual might not be able to recall, an emergent property. Emergent behavior occurs when groups can do things that individuals cannot. Ant and bee communities are examples of emergence where intelligence arises out of a multiplicity of relatively simple and seemingly unmotivated actions. No single ant “knows” that the hill needs to relocate, for example, but the actions of tens of thousands of ants result in the hill being moved, efficiently, effectively, even “intelligently.” The Stanford biologist Deborah Gordon writes, “The basic mystery about ant colonies is that there is no management.” No ant stands at the periphery of the colony directing traffic.: “Hey you! Stop playing ‘rub-the-feelers’ with that worker ant and get a move-on! Come on guys, break it up, there’s enough moldy peanut for every—HEY BUDDY! What’re you, taking a break? Go help those guys carrying the heavy praying mantis carcass!” With no ant in charge, how on earth do ants get anything done?
Ant colonies exhibit behavior very similar to that of other systems with a very large number of units or components, all of which interact, and the consequences of whose interactions change over time. Physicists call these nonlinear dynamical systems. (They’re called nonlinear because the effects of these interactions can’t simply be added up, they sometimes have to be expressed as powers, or other higher mathematical functions. They’re called dynamical because the influence of one event at the beginning can have profound effects later in time as that initial effect is carried forward.) In systems like these—which include rain forests, stellar transits, the stock market, and even the faddish propagation of hit songs—small, seemingly chaotic, and unrelated behaviors can end up having large effects as they interact, spread, and develop over time. In other words, extraodinarily simple individual units—like ants, neurons, atoms, or musical notes—can generate complicated and often counterintuitive global behavior.
The World in Six Songs: How the Musical Brain Created Human Nature Page 17