The Note Is Where the Bend Ends

You can learn the right notes of a guitar phrase and still miss the phrase.

That sounds like the usual sermon about feeling: fewer notes, more soul, stop staring at the fretboard. But there is a smaller, more useful problem hiding inside it. Sometimes the thing missing is not mysterious expression. It is information you threw away when you decided what the notes were.

Consider a phrase Kirk Fletcher uses in a B.B. King-style lesson. In A, Fletcher describes E–F#–A, followed by B bent a whole step to C#, then E. He presents it as a teaching example along the lines of the introduction to “Sweet Little Angel,” not as an exact transcription of a particular King recording.[1]

Write only the destinations and you get:

E–F#–A–C#–E.

Keep the gesture and you get:

E–F#–A–[B bent to C#]–E.

Those lists end in the same places. They do not describe the same journey. C# is the arrival; B is where that particular bend begins. A direct C# attack cannot preserve a B-to-C# approach that it never makes.

This is what changed my understanding of learning a solo by ear. The job is not simply to replace tablature with better ears. It is to decide what musical information you are trying to recover—and to notice which parts of it your representation leaves out.

Three jobs pretending to be one

“Learn that solo” bundles together at least three tasks.

Reconstruction: form a musical hypothesis about what happened. What pitch? In which register? How did it arrive? Where did the phrase sit against the beat? Which details are clear, and which are guesses?

Reproduction: make that hypothesis happen on your instrument. Find a position, a fingering, a picking or legato solution, a bend you can actually control. Your fingers can disagree with your understanding even when the understanding is sound.

Transfer: take something you learned into a different context. Not necessarily the entire solo. Perhaps a contour, a way of delaying an arrival, or a short phrase whose relationship to the harmony you understand.

These are working distinctions, not three sealed stages. Playing can expose a bad listening hypothesis. Listening again can reveal that your elegant fingering solves the wrong problem. But keeping the jobs distinct makes a mistake easier to locate. “I can't play it” might mean “I haven't worked out what it is,” “I know what it is but can't execute it,” or “I can execute it there and don't know where else it belongs.”

Michael Kaplan's transcription advice offers one route through that tangle. In a Berklee interview, he describes internally hearing and singing a solo, then finding it on guitar, playing from memory, and notating it. He also recommends extracting phrases and exploring keys and fretboard positions.[2] This is practitioner guidance, not experimental proof that one sequence is best for every learner. Its useful idea is the separation: first establish something you can hear internally; don't let the instrument's familiar shapes quietly decide the whole answer.

The bend contains more than its interval

The Fletcher example lets us make one exact claim and stop before pretending to know more.

Its published prose specifies a whole-step bend from B to C#. That is enough to distinguish a pitch route from a destination. It is not enough to reconstruct the bend's speed, its onset relative to the beat, how long it waits before arriving, or the sound of its release. Those are additional questions, not details I can recover by enthusiasm.

I have not inspected the lesson audio or performed the phrase. This is a study of an attributed written example, not a claim that I transcribed King—or even heard Fletcher's demonstration. The missing evidence matters because the subject is listening. It would be especially absurd to write an article about preserving musical information while inventing the information I never encountered.

For a guitarist working with an actual recording, the corresponding question is concrete: What would I need to hear again before I could defend this version? Perhaps the bend is clear but its timing is not. Perhaps a second instrument masks the attack. Perhaps the pitch is certain and the articulation remains provisional. A useful transcription can retain those differences instead of announcing one undifferentiated “correct.”

Tablature is not the villain

It is tempting to make this a morality play. By ear is authentic; written notes are cheating. That is too easy.

Kaplan explicitly allows written transcriptions as scaffolding when a learner cannot yet recover the material by ear. His insistence is that listening still matters for feel, articulation, and nuances.[2] The distinction is not paper versus purity. It is which decisions you made, which decisions someone supplied, and which ones you checked against the sound.

A destination-only list loses the bend. A carefully marked transcription can preserve it. An impressive display of ear learning can still miss it. The format alone does not tell us whether the gesture survived.

There is some research relevant to the separation between hearing and doing, but it needs a smaller claim than “science proves you should learn every solo by ear.” In a 2010 study, Robert Woody and Andreas Lehmann compared melody reproduction by singing and by instrument among 24 participants with different musical backgrounds. The author-provided abstract reports fewer trials for singing than instrumental reproduction and differences associated with vernacular versus formal musical experience.[3]

The abstract reports background comparisons, not a randomized ear-versus-tablature intervention. It does not establish that transcribing a solo produces better improvisational transfer. My access here is to the abstract, not a full methods or effect-size analysis. It supports taking encoding and instrumental realization seriously as different parts of ear playing. It does not settle the pedagogy contest.

Move the relation, not just the shape

The Fletcher phrase also makes transfer less vague.

Transpose every pitch upward by five semitones, from an A-centered example to D, and the written gesture becomes:

A–B–D–[E bent to F#]–A.

The bend is still a whole step. Its arrival is still the major third of the named center: C# in A, F# in D. That is a checkable relationship, not a claim about how well anyone plays it. A guitarist could reproduce the relationship in another position—or preserve the fingering while accidentally landing on something else.

Now change more than the key. Put the idea in a D-minor context. In D minor, the tonic chord's third is F, not F#. That does not make F# universally forbidden. It means the same arrival has a different relationship to the context. Before retaining, changing, or rejecting it, you have to decide what you want it to do.

So “I can move this lick” has at least two meanings. You can relocate its pitches while preserving their relationships, or you can reconsider those relationships in a changed harmonic setting. The first is already useful. The second asks more of you.

Here is a modest practice proposal—not a session I have done, and not a validated training protocol:

Choose a short phrase from a recording you can revisit. Recover more than its destination notes; mark one uncertainty rather than hiding it.
Try to sing or otherwise hold the phrase internally, then reproduce it on the instrument. Notice whether the sticking point is the musical hypothesis or its execution.
Change one condition: key, position, or harmonic setting. Say what you mean to preserve, and what you deliberately change.
Return to the larger passage. A fragment can survive extraction while losing its role in the solo.

The experiment is not successful merely because the notes fit a diagram. It needs listening and playing. The diagram helps name what those actions should test.

Imitation need not be the opposite of invention

A final complication is worth keeping. Copying and creativity are not necessarily rival camps.

Emily Sayers's 2025 study of North Indian classical music pedagogy describes imitation, memorization, and structured variation or recomposition as parts of developing improvisation and composition.[4] That observational work gives us a counterexample to the idea that imitation must stop before invention can begin. It does not prove that the same teaching practices transfer unchanged to blues guitar, or that a particular amount of copying guarantees originality.

For this question, the useful distinction is more modest: imitation can preserve material; variation asks which relations are available to be reworked. Remembering the lick and understanding its possibilities can feed each other without becoming the same accomplishment.

I started with “learning a solo by ear” as if the phrase named one skill. I now see a chain of recoveries and decisions. What happened? Can I realize it? What part of it do I understand well enough to carry elsewhere?

The next interesting question is not how much of a solo someone can memorize. It is what survives when they move one phrase—and whether they can hear what they left behind.

Sources

[1] Kirk Fletcher, “How to play the blues like B.B. King,” Guitar World, August 12, 2022. Figure 1 explanatory paragraph. An attributed King-style teaching example, not an exact King-recording transcription. https://www.guitarworld.com/lessons/how-to-play-like-bb-king

[2] Jonathan Feist interviewing Michael Kaplan, Berklee Online Take Note, “Transcribing Jazz Guitar with Michael Kaplan.” Practitioner recommendations; especially discussion of written versus ear transcription, workflow, and using phrases in performance/practice. https://online.berklee.edu/takenote/transcribing-jazz-guitar-an-interview-with-michael-kaplan/

[3] Robert H. Woody and Andreas C. Lehmann, “Student Musicians' Ear-Playing Ability as a Function of Vernacular Music Experiences,” Journal of Research in Music Education 58(2), 101–115, 2010. Author-provided abstract only, ERIC EJ889094. https://eric.ed.gov/?id=EJ889094

[4] Emily Sayers, “Pedagogical strategies for the development of improvisation and composition in North Indian classical music,” Frontiers in Psychology 16:1460158, June 4, 2025. Abstract used for the limited observational comparison. https://doi.org/10.3389/fpsyg.2025.1460158