The Theory-Practice Gap: Why the Grid Is Not the Music

Modern theory gives us elegant grids for makam and dastgah: 53 Holdrian commas per octave for Turkish music, 24 quarter tones for Arabic, interval tables measured to the cent. These frameworks are genuinely useful — they let us notate, teach, compare, and build software for music that was transmitted purely by ear for centuries. But they are maps, not the territory. Every serious empirical study of how performers actually play this music reveals the same uncomfortable truth: the grid does not match the practice.

This article examines what we know about that gap — from spectral analysis to computational pitch studies to the testimony of the traditions themselves — and what it means for anyone trying to learn makam or dastgah music today.

The promise of theoretical frameworks

Established The Arel-Ezgi-Uzdilek (AEU) system, adopted by Turkish conservatories in the 1930s, divides the octave into 53 equal Holdrian commas. Each perde — each named pitch position — sits at a precisely specified comma value. A whole tone is 9 commas. A cesni (tetrachord or pentachord) is defined as a specific sequence of comma intervals. The system is internally consistent, mathematically clean, and powerful enough to express the interval relationships of every standard makam.

Established The Persian tradition never adopted an equivalent grid. Farhat, in The Dastgah Concept in Persian Music, explicitly rejected fixed interval systems for Persian classical music, arguing that the dastgah system resists reduction to any equal-temperament framework. But Persian music still has theoretical descriptions — interval ratios, scale diagrams, forud patterns — that create their own kind of map.

Practice Both maps are useful. Neither is the music.

What performers actually do: Akkoc's non-deterministic scales

Attested In 2002, Can Akkoc published a study that reframed how we think about pitch in Turkish makam performance. In "Non-Deterministic Scales Used in Traditional Turkish Music" (Journal of New Music Research, 2002), Akkoc analyzed pitch data from recordings of master performers and argued that the pitches they produce are not fixed points but probability distributions — clouds of pitch around target positions, not single comma values.

Attested The central finding was this: when a performer plays a given perde repeatedly, the resulting pitches do not cluster at one exact frequency. They form a distribution — a spread of frequencies centered roughly near the theoretical position but varying from instance to instance, from phrase to phrase, from performer to performer. Akkoc argued that this variability is not error or imprecision. It is a fundamental feature of the music — a non-deterministic system that cannot be adequately described by any fixed scale, whether 53-TET or otherwise.

Practice This finding has significant implications. If a perde is not a point but a region, then the comma grid is a useful approximation that should not be confused with a specification. A student who practices hitting exact comma positions is practicing something, but it is not quite what the masters do.

Computational confirmation: Bozkurt's pitch histograms

Established Barış Bozkurt and his collaborators brought computational methods to bear on the same question. In a series of studies beginning with "An Automatic Pitch Analysis Method for Turkish Maqam Music" (Journal of New Music Research, 2008) and continuing with "Pitch-Frequency Histogram-Based Music Information Retrieval for Turkish Music" (with Ali C. Gedik, Signal Processing, 2010), Bozkurt developed techniques for extracting pitch histograms from large corpora of makam recordings.

Established Bozkurt's method uses the Holdrian comma as a unit of measurement — the pitch axis of his histograms is divided into commas — but the histograms themselves reveal that performed pitches do not sit precisely at theoretical comma positions. Peaks in the histogram are often offset from the AEU-predicted values, and the width of those peaks confirms the distributional character that Akkoc described. Different performers, different instruments, different recordings of the same makam yield pitch distributions that overlap but do not perfectly coincide.

Attested Bozkurt's work also demonstrated that despite this variability, pitch histograms carry enough consistent structure to support automatic makam recognition and tonic detection. The patterns are real — they are just softer, wider, and more context-dependent than theoretical comma positions suggest. The music has structure; it is simply not the structure the theory writes down.

Farhat and the dastgah system: flexibility as doctrine

Established In the Persian tradition, the gap between fixed theory and living practice is not hidden — it is, in a sense, officially acknowledged. Farhat's The Dastgah Concept in Persian Music documents the dastgah system as fundamentally resistant to rigid pitch specification. The concept of moteghayyer — the mutable pitch, a note that consistently appears at two distinct frequencies depending on melodic context and the performer's expressive intent — is built into the theory itself.

Established The radif, the canonical body of melodic models (gusheh) that defines each dastgah, is transmitted as a repertoire of melodic shapes, not as a scale. A gusheh specifies which pitches receive emphasis (shahed), where cadential descents (forud) land, and which pitches may shift (moteghayyer) — but it does not pin those pitches to exact frequency values. The performer learns the melodic identity through repeated practice with a master, not from a tuning chart.

Practice This is arguably more honest than the Turkish theoretical apparatus. The AEU system implies a precision that performance does not deliver. The dastgah tradition names the flexibility and builds it into the pedagogy. But both traditions arrive at the same place in practice: the performer's ear, trained by long apprenticeship, is the final authority on pitch.

The meşk tradition: what notation displaced

Established Before conservatories, Turkish classical music was transmitted through meşk — a face-to-face, master-to-student oral tradition documented extensively by Feldman in Music of the Ottoman Court. In meşk, the student learned compositions and makam behavior by performing alongside a teacher, internalizing pitch relationships, seyir patterns, ornamental conventions, and expressive timing through imitation and repetition over years. There was no score to follow. The music lived in the teacher's memory and the student's ear.

Attested When Turkey's music conservatories adopted Western-style staff notation augmented with microtonal accidentals in the early Republican period, something was inevitably lost. Notation captures pitch as a fixed value — a specific line or space, modified by a specific accidental. It cannot represent the distributional, context-dependent nature of pitch that meşk transmitted implicitly. A student who learns Makam Ussak from a score sees a sequence of fixed perdeler. A student who learned it through meşk absorbed a melodic behavior — including which pitches flex, in which direction, under which expressive conditions.

Attested Signell, writing in 1977, already noted discrepancies between what the AEU system prescribed and what performers actually played. These discrepancies are not evidence of sloppy playing. They are the residue of a practice tradition that predates — and in some ways exceeds — the notational framework imposed on it.

Why this matters for learners

Practice If you are learning makam or dastgah music, the theory-practice gap is not an abstract concern. It shapes how you should practice:

1. Learn the grid, but don't worship it. The Holdrian comma system, interval tables, and scale diagrams are essential for building initial understanding. They give you landmarks. But they are starting positions, not final destinations.

2. Train your ear on recordings, not just theory. The pitch relationships you need to internalize live in the performances of masters, not in textbook interval tables. Listen to how a perde shifts in different melodic contexts — that shift is not error, it is the music working as intended.

3. Understand seyir as the real identity of a makam. The melodic path — which pitches receive emphasis, where phrases rise and fall, how the makam unfolds over time — matters more than the exact cent-value of each perde. Two performers may disagree on the precise intonation of segah by 15 cents and still agree completely on the seyir of Makam Segah.

4. Respect moteghayyer. In dastgah music, certain pitches are defined as mutable. This is not a theoretical curiosity — it is an instruction. When you encounter a moteghayyer in a gusheh, you are expected to vary it. Fixing it at one frequency misses the point.

5. Seek out oral transmission. If you have access to a teacher in the tradition, the face-to-face transmission of melodic behavior — even informally — conveys things that no score, recording, or article (including this one) can fully capture.

The music is living

Practice The theory-practice gap is not a bug. It is evidence that makam and dastgah are living musical systems, not mechanical ones. A fixed grid would make these traditions easier to digitize, easier to teach at scale, easier to build software for. But it would also flatten them — reducing a rich, context-dependent, expressively flexible art to a set of MIDI pitches.

The frameworks we use — 53-TET, comma tables, interval diagrams — are scaffolding. They help us approach the music, talk about it, compare traditions. But the music itself lives in the space between the grid lines: in the slight sharpening of a perde as a phrase ascends, in the moteghayyer that tilts one way in a daramad and another way in a forud, in the seyir that a performer has absorbed so deeply that it feels like instinct rather than theory.

Learn the grid. Then learn to hear past it.