[DSRP Evidence](https://dsrpevidence.org/)

# Temporal Gestalt Perception in Music

## Details

**Authors** Tenney, James and Polansky, Larry

**Year** 1980

**Publisher** Journal of Music Theory

**Kind of work** article

**Discipline** Music Theory

**Secondary disciplines** Cognitive Science, Computer Science & AI

**Applied** false

[Read it at the publisher](https://doi.org/10.2307/843503) 
10.2307/843503

## In authors' words

### What they set out to do (purpose)

To ask whether the perceptual boundaries of musical temporal gestalt-units can be predicted and simulated by a computer algorithm.

### Who or what was studied (sample)

Four monophonic works: Varèse's Density 21.5, the second movement of Webern's Concerto Op. 24, Debussy's Syrinx, and the English horn solo from Wagner's Tristan und Isolde.

### How they did it (methods)

Music is represented as points in a multidimensional space of time, pitch and intensity, with city-block distance between successive elements. A new unit is initiated where the distance from the preceding element exceeds the distances immediately before and after it. At higher levels a disjunction measure combines mean-distances and boundary-distances, with weights that halve at each lower level. Weightings were tuned piece by piece.

### What they found (results)

The program locates unit boundaries at several hierarchical levels (element, clang, sequence, segment, section) that the authors describe as at least plausible. Best weightings differed from piece to piece (equal weights for Debussy; pitch strongly suppressed, 0.03, for Wagner) and were found by trial and error. The program identifies boundaries but says nothing about their musical function.

## Commentary

### In short

The authors build a computer procedure that divides a melody into nested units, from single notes up to whole sections, by finding where the music jumps most in time, pitch and loudness. Applied to four solo pieces, it produces plausible segmentations.

### Response

The segmentation depends on a boundary rule (distinction) that separates each unit from its neighbors and on units nested inside larger units (system), with each level requiring at least two units from the level below. Caveats: the work models only monophonic music, the weights were fitted by hand per piece, and the authors give no independent test of the boundaries; the fit to relationships and perspectives is not argued here.

**Patterns it shows** D, S

Shows the structure is there

**Added** 2026-10-26

**How to cite this** Tenney, James and Polansky, Larry (1980). Temporal Gestalt Perception in Music. Journal of Music Theory.