A Disorganized Rant About Music Literacy

I’m reading a very interesting facebook thread right now about music literacy. Here is my response to some of the ideas on that thread.

_________________________________________________________________________________________

Dr. Gordon wasn’t perfect. He made several notable mistakes, usually when he swung too far in the opposite direction of a problem. Teaching students to recognize symbols, and calling that reading, is a problem; naming notes by use of gimmicks such as “Every good boy does fine,” and calling that reading, is a problem. Gordon sought to fix these deficiencies in music education, which he described as “decoding” notation, but his solutions, in my opinion, were less than stellar.

By the way, Gordon was wrong in calling note-naming “decoding.” Decoding is a good thing; decoding is the same as notational audiation. What Gordon objected to was recoding, which Goodman (1968) described as “sounding out” print, but not attending to meaning. Gordon also objected to symbolic recognition — merely naming symbols but not sounding out or understanding their meaning. I agree with these sentiments. But Gordon should not have used the word decoding: When you decode a coded message, you understand it; and when you paraphrase the meaning of a coded message, you understand it deeply.

Decoding is the good guy; symbolic recognition is the bad guy.

And what about recoding — the ability to convert notation to sound? Recoding is a good guy too, but it’s incomplete. Recoding notation — “sounding it out,” in other words — means nothing if students don’t understand the syntax of what they’re reading. Recoding is like reliability in test design: it’s necessary but not sufficient.

The psycholinguist Frank Smith (1994) pointed out something similar in language reading: just being able to sound out words, or recognize whole words, does not mean you understand the word, the sentence, the paragraph, the piece. But in his denunciation of phonics, Smith went too far. More than 100 experimental, randomized-controlled studies over the past 100 years have shown that phonics is indispensable for beginning readers. Chall’s (1996a) exhaustive book Learning to Read: the Great Debate, in which she reviewed 87 studies from 1912 to 1967, should have settled the matter. Her 1996 updated edition of that book should really have settled the matter. The meta-analysis put out by the National Reading Panel (2000) should have put the final nail in the whole-word/whole language coffin. But no. In every generation, people keep reinventing anti-phonics nonsense.

I’m not saying phonics is enough. In her brilliant book Stages of Reading Development, Chall (1996b) theorized that phonics is both a level of learning (with many sub-levels) and one of 6 stages of reading development. Her theory, as I see it, may be applied to music reading as well: symbolic association is a level of learning and a stage of music reading development.

Composite synthesis, by the way, is not a level of learning, but is, in my opinion, a stage of music reading development. In my book (2000), I think I got that wrong. But that’s for another blogpost.

I once asked Dr. Gordon who he thought his most important influences were, outside of music, when he constructed his skill-learning sequence. He unhesitatingly said that the two thinkers who influenced his skill-learning sequence most were Frank Smith and Robert Gagne — which struck me as odd because Frank Smith trashed the very idea of sequential learning. I resisted the temptation at that moment to say to Gordon that his thinking was at its best when he learned from Gagne, and at its worst when he followed Frank Smith down the rabbit hole of anti-phonics.

_________________________________________________________________________________________

Do We “Bring Meaning” To Notation?

Frank Smith and Kenneth Goodman, two prominent psycholinguists in the 1970s and 1980s, held the wrong-headed opinion that we “bring meaning” to notation — an idea MLTers are saddled with to this day. In a New York Times Interview from 1973, journalist Edward Fiske asked Goodman the following question: If the sentence is “The boy jumped on the pony and rode off,” and a student reads it as “The boy jumped on the horse and rode off,” should the teacher correct him? No, the teacher should not correct him, according to Goodman. The student brought basically the correct meaning to the sentence. And that’s good enough.

Except it’s not good enough! A horse is not a pony!

And this is the problem I have with the “bringing meaning to notation” business: accuracy is devalued.

Skilled readers bring to notation not meaning but understanding. Of course such understanding may vary from one music reader to another, especially when it comes to phrasing, style, and general interpretation. And if the music is tonally and metrically ambiguous, the meaning of the notation gets even murkier. But even in such a case, there are limits to how valid your interpretation can be.

Don’t get me wrong. I love Gordon’s idea that we audiate the essentialness of a piece of music; I love the idea that we may disagree about which pitches and durations are essential. In fact, I’m writing a blogpost now called, “The Most Ignored But Essential Part of MLT is … Essentialness.”

_________________________________________________________________________________________

Sounding Out Notation

Gordon seemed to suggest — he never said it outright — that students who have learned to read and audiate familiar patterns can look at music notation, familiar or not, and then audiate the essential pitches and durations effortlessly. But there’s an intermediate step he never talked or wrote about: Before you can decide what pitches and durations are essential in a notated piece of music, you must “sound out” every, every, every notated pitch and duration.

Maybe you doubt this. If so, think about regular, non-notational audiation for just a minute. You hear a piece; then you go to work, sorting out the essential pitches and durations; with those building blocks, you form patterns, you audiate a resting tone and macro and micro beats, then you audiate tonality and meter, and so on. You can do this because you have all the pitches and durations to work with.

As I see it, notational audiation works basically the same way, but with one important addition: To read and understand music, you must sound out the notation, that is, you must convert written notes into sound. (The music may be outwardly or inwardly heard; it doesn’t matter.) Once you’ve sounded out the written pitches and durations — all of them, essential or not — then, and only then, can you audiate the notation accurately. In short, music readers cannot bypass sound to get to meaning.

TRADITIONAL MLTer: We don’t bypass sound. We teach students to read/sing familiar patterns; and then they can read/sing unfamiliar patterns.”

ME: But can they? Are students really able to read/sing unfamiliar patterns after singing whole, familiar patterns at sight? What tools have you given them?

TRADITIONAL MLTer: Once they learn to read familiar tonic and dominant patterns in major, then they can notationally audiate unfamiliar patterns with the same functions.

ME: Yes, students can audiate the unfamiliar patterns they see, but only if someone else reads the patterns for them. The question before the house is, What skills do MLT-trained students have to get the notes off the page before they audiate the music they’re reading? None. MLT fails to provide students with crucial recoding skills.

TRADITIONAL MLTer: For students to read unfamiliar music successfully, they don’t have to learn to recode directly. Students will do this naturally. ‘Sounding out’ notes is not a skill we need to teach.”

The traditional MLTer may have a point. Perhaps recoding is not a skill we must teach our students directly; maybe students will pick it up on their own.

I built my dissertation study on the notion that students must have the tools to sound out unfamiliar written music (recode it, in other words) before they can audiate it. This was the central question of my study: Can students teach themselves to convert written notes to the pitches those notes represent, or must they be taught to do so directly? If you’re interested in learning more about this, please read the post Music/Language Analogies Part 6: Phonemes and Morphemes found here.

_________________________________________________________________________________________

Coda: A Few More Thoughts About “Bringing Meaning”

To finish this post, I’ll rant a bit more about “bringing” meaning to notation. Let’s say you’re reading a piece by Mozart, who has written an unambiguous IV-V-I progression; and let’s say you “bring” the meaning to it that Mozart actually wrote a ii-V-I progression. Sure, you can do that. Hell, it’s a free country, so therefore, you can “bring” any meaning you damn well please.

But let me say with all the strength I can muster that you have a responsibility — to Mozart, to yourself, to the art of music — to read music accurately. And how do you do this? By bringing recoding skill to the notes Mozart wrote. And once you recode Mozart’s notation, you can take the next crucial step: you can audiate (in other words, decode) the notation.

And what about bringing meaning? Mozart’s notation is so saturated with meaning that to bring meaning to it is not only unnecessary but arrogant. We can only humbly take meaning from it, by bringing understanding to it.

The phrase “bring meaning to notation” suggests that it’s ok to bring any meaning you want to what you see — a ii-V-I instead of a IV-V-I progression. But you pay a price for bringing the wrong meaning to Mozart’s notation; and the price you pay is that you’re wrong. Despite what Goodman said in his New York Times interview, there’s nothing good to be said about being wrong.

_________________________________________________________________________________________

Bluestine, E.  (2000).  The ways children learn music:  An introduction and practical guide to music learning theory.  Chicago:  GIA.

Chall, J. S.  (1996a).  Learning to read:  The great debate (3rd ed.).  New York:  McGraw-Hill.

Chall, J. S.  (1996b).  Stages of reading development (2nd ed.).  New York:  McGraw-Hill.

Fiske, Edward B. “Approach to Reading Rethought” (interview with Kenneth Goodman). New York Times, July 9, 1973.

Goodman, K.  (1968).  The psycholinguistic nature of the reading process.  In K. Goodman (Ed.), The psycholinguistic nature of the reading process (pp. 15–26).  Detroit:  Wayne State University Press.

National Reading Panel (NRP).  (2000).  Report of the National Reading Panel:  Teaching children to read:  An evidence-based assessment of the scientific research literature on reading and its implications for reading instruction: Reports of the subgroups.  Rockville, MD:  National Institute of Child Health and Human Development Clearinghouse.

Smith, F.  (1994).  Understanding reading:  A psycholinguistic analysis of reading and learning to read.  Hillsdale:  Lawrence Erlbaum Associates.

Meter Test — Second Revision

Here is the latest version of a Meter Discrimination listening test I began writing many years ago. I constructed this test before Covid, administered it once, threw out poorly discriminating items, revised it, shortened it, administered to a new group, got decent difficulty and discrimination numbers, shortened it some more, reordered the items to spread the difficult items around; and in 2020, I was all set to administer it again to a new group of students so I could obtain reliability estimates. Then Covid struck.

I’m using the test this year with my 4th graders to fulfill a Philadelphia School District requirement: their SPM (Student Performance Measure). To meet my school districts expectations, I must treat the test as a criterion-reference test (basically a pass/fail mastery test). Actually, it’s more a test of my teaching than of my students’ learning. Here is how it works (and now, I’m quoting from the Philadelphia School District SPM template):

To achieve an Educator Effectiveness rating of Distinguished, 80% of students will score a 19 or above on the post‐test assessment.

To achieve an Educator Effectiveness rating of Proficient, 70‐79% of students will score 19 or above on the post‐test assessment.

To achieve an Educator Effectiveness rating of Needs Improvement, 60‐69% of students will score 19 or above on the post‐test assessment.

To achieve an Educator Effectiveness rating of Failing, less than 60% of students will score 19 or above on the post‐test assessment.

_________________________________________________________________________________________

Here are the test item sound files:

1. Bizet – Carmen – Overture

2. Schubert – German Dance in B-flat, D. 783 #7

3. The Beatles – Norwegian Wood

4. Scarlatti – Sonata in G Major, K. 455 (synthesizer)

5. Bernstein – West Side Story – I Feel Pretty

6. Prokofiev – Classical Symphony, 3rd mvmt

7. Seal – Kiss from a Rose

8. Beethoven – Symphony #9, 4th mvmt

9. The Del-Vikings – Come Go with Me

10. Johann Strauss – Tales from the Vienna Woods

11. Verdi – Rigoletto_ La donna è mobile

12. Mozart – Eine Kleine Nachtmusik 1st mvmt

13. James Taylor – Sweet Baby James

14. Johann Strauss – Blue Danube Waltz

15. Handel – Parnasso In Festa – Chorus

16. Haydn – Symphony #103, Drumroll– III. Menuet

17. Verdi – Il Trovatore – Anvil Chorus

18. Bach – Toccata In D, BWV 912

19. Billy Joel – Piano Man

20. Banjo solo – Cherokee Shuffle

21. Handel – Brockes Passion – Gott selbst der Brunnquell

22. Bizet – Carmen Suite No. 1_ VI. Les toréadors

23. Smetana – Bartered Bride – Act 1, Scene 1

24. Molly Mason & Jay Ungar – The Lovers’ Waltz

25. Bizet – Carmen Suite No. 1- IV. Seguidilla

26. The Marcels – Blue Moon

27. Handel – Water Music, Suite #2 – Minuet

28. Gotye – Somebody That I Used to Know

29. Saint-Saëns – Carnival of the Animals – Fossils

30. Kelly Clarkson – Breakaway

31. Saint-Saëns- Samson et Dalila – Bacchanale

32. Rankin Family – Mo Shall al dhekh

_________________________________________________________________________________________

Please take a look at FIGURE 1 below. 

FIGURE 1

You’ll see that I present item analyses for all items except 22 and 24. For this revised version of the test, I replaced those items with different musical examples by the same composer/performers (Bizet and Mason & Ungar respectively). I ordered the items with 4 factors in mind: meter, genre, item difficulty, and item discrimination. Let me go into this in greater detail.

_________________________________________________________________________________________

Technical Matters

—In this 2nd revised version, the meter test consists of 32 items, 16 in duple meter and 16 in triple meter. I threw out 8 test items (4 in duple meter and 4 in triple meter) from the original test because of their poor discrimination values.

—I shortened the test items.  Each item is now between 25 and 30 seconds long.  Most are exactly 27 seconds long. (Incidentally, I have no way of knowing, yet, if shortening the items will have any effect on how students respond to them.)

—No more than 2 duple examples appear consecutively, and no more than 2 triple examples appear consecutively.  If students circle duple or triple only, or if they consistently alternate their responses, they will receive a chance score of 16.

—The revised test has 15 vocal items (shown in bold print) and 17 instrumental items. You’ll notice that pieces similar in style and genre are not lumped together.  No two vocal items appear consecutively; in a few instances, two instrumental items appear consecutively (#1 and #2; #24 and #25).  This was inevitable.  Still, I made sure the back-to-back instrumental items are stylistically different from each other.

—The 4 very difficult items (shown in green font) are those with difficulty levels at 60% or lower).  I deliberately spaced them apart, so students will not face the daunting task of answering many difficult items in a row.

—The 3 very easy items (shown in red font) are those with a difficulty level of 90%.  I deliberately spaced them apart so kids, while taking the test, will get a periodic breather.

— I learned in a previous administration of the test that half the test items have difficulty levels at or above 75%, while half have difficulty levels below 75%.  If you look at the item difficulty levels, you may think the items were randomly scattered, so that the varied difficulty levels happened by chance. Actually, to make sure items varied in difficulty levels throughout the test, I arranged them by following this procedure: I divided the 32 items into four 8-item groups based on item difficulty. I then made sure each 8-item group had 4 above-average items and 4 below-average items. These groups then became items 1-8, 9-16, 17-24, and 25-32.

—Based on a previous administration of the test, all 32 items are positively discriminating, and only 3 have discrimination values lower than .2. (This does not include the new items 22 and 24.)

In Figure 2 below, you’ll see the format of the test.

FIGURE 2.

_________________________________________________________________________________________

Final Technical Points and Some Personal Reflection

In September 2023, I administered this Meter Test as a pre-test to 4th grade students (n = 103) at the Stephen Decatur School in Philadelphia. As shown in Figure 3, the obtained pre-test mean of 15.07 is encouragingly low, and in fact, is a full point lower than the chance score of 16. The standard deviation (not shown in Figure 3 but calculated separately by use of an excel spreadsheet) is 3.68, which means the scores were all over the map, and basically the kids had no clue what they were doing. Which is exactly what a pretest should reveal!

FIGURE 3.

I just listened to the whole test all the way through, and I’m happy with the sum of its parts.  I got rid of some of the gentler items, so the revised test sounds a bit more raucous than the original version; and because the items are shorter, the whole test moves at a faster pace. (From start to finish, the test clocks in at 14 minutes.) But I’m glad it still has moments of humor, like the “Blue Moon” example that kids love, and Glenn Gould’s positively goofy performance of the Bach Toccata.

In May 2024, following 7 months of instruction in musical meter discrimination, I will administer the test again. My predictions are that the same group of students who took the pretest will obtain the theoretical mean of 24 and a theoretical standard deviation of 2.67. After I calculate the mean and standard deviation, I will analyze each item to determine difficulty levels and discrimination values; and finally, I will obtain test-retest and Kuder-Richardson 20 reliability estimates.

And now, we come to the hard part: waiting. Will the items retain their varied difficulty levels and their high discrimination values?  And what about the reliability of the test?  Will students’ scores be relatively stable over time?  And will the test show a high degree of internal consistency?  By next May, I should have the answers.  As Willy Wonka said, “The suspense is terrible.  I hope it’ll last.”