the interplay between metric and figural rhythmic organization

16
Journal of Expert mental  Psychology: Human  Perception  and Performance 1998,  Vol.  24, No.  5,1546-1561 Copyright  by  the American Psychological Association, 0096-1523/9WS3.00 The Interplay Between Metric and Figural Rhythmic Organization Stephen Handel University of  Tennessee,  Knoxville Three experiments investig ated  the relative  importance  of figural and  metric rhythmic organizations. Figural organization  is  determined  by the  numbers  of  tones  in  successive groups.  Fo r  figural  organization alone,  th e  timings between  th e  onsets  of  each group  are relatively unavailable, so listene rs cannot disc riminate between 2 rhythms that have the same sequence  of  groups  but  different  timings between  th e  groups. Thus, traditional views argue that  a  metric organization  is  necessary:  The  timing between adj acent groups  is  perceived  b y means  of the  strong-weak  sequence  of  beats.  These  experiments, however, sugg est  a  limited role  fo r  meter.  Th e  metric strength  of the  individual rhythms  affected  discrimination  of  pairs  o f different  rhythms  with  th e  same  figural  organization only when  an  external meter pulse accompanied  th e  rhythm  a nd  only when  the  rhythm with  th e  stronger meter  was the first of the pair. Th e  purpose  of the  present experiments  was to  investigate th e  roles  of the  metric  (i.e.,  periodic)  and fig ural  (i.e., grouping or serial) organizations in the perception of rhythmic  timing patterns composed  of  identical tones. Traditionally, rhythmic organization  h as  been thought  to depend on identification of the meter or beat of the pas sag e. Passages  in which s uch a meter cannot be easily cr eated are considered  to be  unstable  and  therefore  difficult  to  encode and  reproduce. The results of the  present experiments argue that  this  is an  overly simplified conceptualization  and that the  metric organization functions only within  th e more fundamental figural organization created by element grouping. Metric  organization  is the  sense  of a  regular periodic sequence of subjectively stronger and weaker beats that ch aracteriz e musi c. The meter forms a time- bas ed lattice that ser ves to create the rhythmic organization. The meter occ urs at  several hierarchical levels  at  once  s o  that  th e  beats  at higher  levels  oc cur at integer multiples of the beats at lower levels.  T he strength of any beat is determined by the number of  levels  at  which  th e  beat appears.  Fo r  example, consider four-beat  meters based  on a  repeating unit  of 16  elements  as used  here.  Th e  strong beats would occu r  at  Elements  1,5,  9 ,  13 ;  stronger beats would occur  at  Elements  1 and 9; and di e  strongest beat woul d occ ur  at  Element  1  (Palmer  & Krumhansl,  1990).  Notes that  fall  at the points of the strong beats become accented,  and  notes  that  fall  at the  points  o f weak  beats are unaccented. Alternative meters could be based on units of three, so that for a repeating unit of  1 2 ,  th e strong est beats would occur on Elements  1,4,7,  and 10 . To  define  strong  and  weak metric rhythms,  I  made  use of th e  rules suggested  by  Povel  an d  Essens (198 5 )  for se- I  thank  Piet  Vos, Greg  Sandell, and  Mari  Riess  Jones  for  their helpful  comments on the manuscript and  Hancel  Woods  for  his help in  completing Experiment  3 . Correspondence concerning this article should  be  addressed  to Stephen Handel, Department  of  Psychology,  University  of  Tennes- see, Knoxville,  Tennessee  37996-0900.  Electronic mail  may be sent  to  [email protected]. quences  of  identical elements separated  by  different  lengths of  silent intervals. Povel and Es sens began by observing that in  a sequence of identical tones, certain tones appear accented. These include  (a )  relatively isolated tones,  (b)  th e second tone  of a  pair  of  tones,  and (c) the first and  last tones in  a  series  of tones. Pov el and Essens continued by arguing that  a strong meter emerges when thos e acc ented tones  fall  at regular  intervals  (i.e.,  beat positions). Empirically (Essens & Povel, 1985),  th e  most important  factor in  determining  th e ease  of  reproducing  th  rhythm  and,  by  inference,  in determining  th e  strength  of the  meter  is  that  a  tone sh ould occur  at the position s of the strong meter beats, and a si lenc e or  rest should  not.  Theoretically,  in  determining  th e  best- fitting  meter,  Povel  and  Essens (1 98 5 ) weighted  th e  lack  o f coincidence  of a  tone  and  beat  as the  most important factor. On  this basis, for rhythms 16 elements long I operation- ally  defined  th e  metric strength  of a  rhythm  by the  occur- rence  of  tones  at the  beat positions  1,  5,9,  and 13.  Thus,  th e rhythm  X.X.X...X...X...  (me Xs  represent tone elements,  th e dots  represent  is ochronous silent time intervals,  and  tones separated  by one  time unit  ar e  heard  as  forming  a  single group)  would  b e  strongly metric, because tones  fall  on the stronger beats  at  Elements  1, 5, 9, and 13, but the  rhythm X.X...X..X.X....  would be only weakly metric because tones fall  only  at  Element  1.  From this perspec tive, rhythms  are not  simply metric  or nonmetric.  Instead, each rhythm  is metric to  some degree, depending  on the  strength  of the meter interpretation  it  evokes. Povel  and  Essens  (1985) argued that li steners attempt  to find a  meter  to fit a  rhythmic pattern  (i.e.,  a  template)  and  that high ly metric rhythms more easily induce  an  internal clock that encode s  th e  rhythm  in terms  of the  meter.  I n  contrast, weak metric rhythms  do not induce  an internal  clock,  and therefore the rhythm cannot be encoded  in  terms  of a tempora l  grid  i n  which every element can  b e  located  and  timed. Figural organization  is the sense that  a  sequence  of  tones is heard as a  series  of discrete groups. Th is organization may be  based  on  shared acoustic properties such  as  duration, pitch, timing, or timbre or may be  based  on a  trajectory  such 1546

Upload: marcelo-c

Post on 03-Jun-2018

220 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 1/16

Journal of Expert mental Psychology:Human Perception and Performance

1998, Vol. 24, No. 5,1546-1561

Copyright  1998 by the American Psychological Association, Inc.0096-1523/9WS3.00

The Interplay Between Metric and Figural Rhythmic Organization

Stephen HandelUniversity of Tennessee, Knoxville

Three experiments investigated  the relative  importance  of figural and  metric rhythmic

organizations. Figural organization  is  determined  by the  numbers  of  tones  in  successive

groups.  For  figural  organization alone,  the  timings between  the  onsets  of  each group  are

relatively unavailable, so listeners cannot discriminate between 2 rhythms that have the same

sequence  of groups but different  timings between  the groups. Thus, traditional views argue

that a metric organization  is necessary: The timing between adjacent groups is perceived by

means of the strong-weak sequence  of beats. These experiments, however, suggest a limited

role for meter. The metric strength of the individual rhythms affected discrimination  of pairs of

different  rhythms  with  the  same  figural  organization only when  an  external meter pulse

accompanied  the rhythm and only when the rhythm with the stronger meter was the first of the

pair.

The purpose of the present experiments was to investigatethe  roles  of the  metric  (i.e.,  periodic)  and figural  (i.e.,

grouping or serial) organizations in the perception of

rhythmic  timing patterns composed  of  identical tones.

Traditionally, rhythmic organization  has  been thought  to

depend on identification of the meter or beat of the passage.

Passages in which such a meter cannot be easily created are

considered to be unstable and therefore  difficult  to  encode

and  reproduce. The results of the present experiments argue

that  this  is an  overly simplified conceptualization  and

that the  metric organization functions only within  the

more fundamental figural organization created by element

grouping.

Metric  organization  is the  sense  of a  regular periodicsequence of subjectively stronger and weaker beats that

characterize music. The meter forms a time-based lattice that

serves to create the rhythmic organization. The meter occurs

at  several hierarchical levels  at  once  so  that the  beats  at

higher levels occur at integer multiples of the beats at lower

levels. The strength of any beat is determined by the number

of levels  at which the beat appears. For example, consider

four-beat meters based on a repeating unit of 16 elements as

used here. The strong beats would occur at Elements 1,5, 9,

and 13; stronger beats would occur at Elements  1 and 9; and

die  strongest beat would occur  at  Element  1 (Palmer  &

Krumhansl, 1990). Notes that fall at the points of the strong

beats become accented,  and notes that  fall  at the points ofweak  beats are unaccented. Alternative meters could be

based on units of three, so that for a repeating unit of 12, the

strongest beats would occur on Elements 1,4,7, and 10.

To  define strong and weak metric rhythms, I made use of

the  rules suggested  by  Povel  and  Essens (1985)  for se-

I thank Piet Vos, Greg Sandell, and Mari Riess Jones for their

helpful comments on the manuscript and Hancel Woods for his help

in completing Experiment 3.

Correspondence concerning this article should be addressed  to

Stephen Handel, Department of Psychology, University of Tennes-

see, Knoxville,  Tennessee 37996-0900.  Electronic mail may be

sent to [email protected].

quences of identical elements separated by different  lengthsof silent intervals. Povel and Essens began by observing that

in  a sequence of identical tones, certain tones appear

accented. These include (a) relatively isolated tones, (b) the

second tone of a pair of tones, and (c) the first and last tones

in a series of tones. Povel and Essens continued by arguing

that a strong meter emerges when those accented tones fall at

regular intervals (i.e., beat positions). Empirically (Essens &

Povel, 1985), the most important factor  in determining the

ease  of  reproducing  the  rhythm  and, by  inference,  in

determining the strength of the meter is that a tone should

occur at the positions of the strong meter beats, and a silence

or  rest should not. Theoretically,  in  determining  the  best-

fitting meter, Povel and Essens (1985) weighted the lack ofcoincidence  of a tone and beat as the most important factor.

On this basis, for rhythms 16 elements long I operation-

ally  defined  the metric strength of a  rhythm by the  occur-

rence of tones at the beat positions 1, 5,9, and 13. Thus, the

rhythm X.X.X...X...X... (me Xs represent tone elements, the

dots  represent  isochronous silent time intervals,  and tones

separated  by one time unit are heard  as  forming  a  single

group) would be strongly metric, because tones  fall on the

stronger beats at Elements  1, 5, 9, and 13, but the rhythm

X.X...X..X.X.... would be only weakly metric because tones

fall  only at Element  1. From this perspective, rhythms are

not  simply metric  or nonmetric.  Instead, each rhythm is

metric to  some degree, depending  on the  strength  of themeter interpretation  it  evokes. Povel  and  Essens  (1985)

argued that listeners attempt to find a meter to fit a rhythmic

pattern (i.e., a template) and that highly metric rhythms more

easily induce an internal clock that encodes  the rhythm in

terms of the meter. In contrast, weak metric rhythms do not

induce an internal clock, and therefore the rhythm cannot be

encoded in terms of a temporal grid in which every element

can  be located and timed.

Figural organization is the sense that a sequence of tones

is heard as a series of discrete groups. This organization may

be based  on  shared acoustic properties such  as  duration,

pitch, timing, or timbre or may be based on a trajectory  such

1546

Page 2: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 2/16

METRIC AND FIGURAL RHYTHMS 1547

as an ascending or descending scale. In the present context,

the grouping is based on the timing between tones, so that

the  groups are composed of tones separated by one silent

intertone interval (one dot in the representation). Bamberger

(1978) and Povel and Essens (1985) termed this  figural

grouping because the groups are figures perceived against anordinal ongoing time.  The  rhythms are organized into

bounded groups of elements that follow one another, but the

tunings between the onsets of the successive groups are not

encoded or compared. For example, the rhythm

X.X..X....X.X... would be coded as 2 tones, silence, 1 tone,

silence, 2 tones, silence (written as 2-1-2-); the lengths of the

silent intervals separating the groups would be coded

roughly, if at all. One would expect a similar rhythm with the

identical figural organization (e.g., X.X....X..X.X..., which is

also coded 2-1-2-) to be easily confused with the  former

rhythm because the timing differences between the groups

are not used.

In previous research (Handel, 1992) I demonstrated thatfor  weak metric rhythms composed  of  identical tones

separated by  different  intertone timings (similar  to  those

discussed above), listeners heard the rhythms in terms of the

groups of tones and could not accurately judge the timings

between groups. When two  different rhythms had the same

figural organization, listeners perceived the two rhythms as

being  identical,  and discrimination was below chance. In

contrast, when two  different  rhythms had  different  figural

organizations, discrimination was quite good.

M y goal in the experiments reported here was to investi-

gate how the figural grouping organization interacts with the

metric organization by using rhythms with stronger as well

as weaker meters. Although all theories of rhythm postulate

that both types of organization jointly determine the emer-

gent  rhythm, the precise relationship between the two is

unspecified. For  example,  Lerdahl and  Jackendoff  (1983)

suggested that the metric organization dominates shorter

sections of a composition but that the grouping organization

dominates longer sections.

If  a strong meter leads to a percept based on a fixed

hierarchical  timing  structure, then that structure should

preserve  the  timings both within and  between groups of

tones. In general, it should be easier to distinguish between

two  different strong metric rhythms or between a strong and

a  weak metric rhythm than between  two  weak metricrhythms. More specifically—and  this is the  focus  of these

experiments—it should be possible to discriminate between

two different rhythms with the same figural organization that

differ in meter. However, if a strong meter does not enhance

rhythmic organization, then the ability to distinguish be-

tween two rhythms should depend on whether their figural

organizations are the same, not on the strength of the meter

of  either rhythm. Moreover,  if the figural  organization  is

primary,  and the metric organization subsequently elabo-

rates the timing of the figural organization, then the effect of

an  imposed external meter pulse should depend on the

specific figural organizations of the two rhythms and should

not improve discrimination among all pairs.

Experiment 1

M y specific purpose in Experiment 1 was to  investigate

discrimination among a wide variety of strong and weak

metric rhythms. The goal was to determine if metric strength

affected  discrimination  of  pairs  of  identical rhythms and

pairs of different rhythms with the same or different  figuralorganization(s).

Method

Participants.  All 57  participants were undergraduates  at the

University  of  Tennessee  who  received  course credit  for  their

participation. They were tested in groups of from 1 to 3.

Rhythms.  All rhythms were based on five tones embedded in a

repeating pattern of 16 grid elements.  The rules used  to construct

the rhythms were as follows: (a) A tone always occurred on the beat

at the first element;  (b) a tone always occurred  on the beat at the

13th element  but never occurred  on the final  three grid elements;

and (c) there was at least one silent element, but no more than three

silent elements, between  any pair of adjacent tones. The  combina-

tion  of  these three rules made  the  longest silent interval three

elements, and one such interval  always occurred  at the end of the

rhythms. However, there could  be  other equally long  silent

intervals, so that organization according to the gap rule, in which

the longest  silent  interval ends  the  rhythm, might be  ambiguous

(Garner, 1974; Handel, 1974). However,  in  practice, this did not

prove to be a problem, and no participant claimed that the rhythms

were transformed  from their starting configuration.

I generated  all of the rhythms using  a  16-element template  for

several reasons. First, the number of possible rhythms mat satisfy

the  above rules is  relatively small,  so it is possible to  adequately

sample the range of rhythmic complexity.  Second,  the rhythms

include simple and complex ones, but no one rhythm is so  difficult

that  it is  impossible  to  pick  up in from two to  four  repetitions.Third, the length of one repetition played at normal tempo is well

within the memory span. Fourth,  a 16-element grid is fit perfectly

by either a two- or four-beat meter. The strongest beats occur at

Elements 1 and 9, weaker beats occur at Elements 5 and 13, and the

weakest beats occur at Elements 3,7,11, and 15. A four-beat meter

seems most natural  to listeners  of Western music: Bolton  (1894)

noticed that listeners spontaneously grouped isochronous tones into

units of 4 and 2 as opposed to 3, and Smith and Cuddy (1989) found

that four-beat meters produced better performance man three-beat

meters.

The preceding rules could generate  19 possible rhythms. Three

rhythms were  perfectly  metric, having tones at Elements 1, 5, 9,

and  13; four rhythms were  strongly  metric,  having tones at

Elements  1, 9, and 13; four rhythms were metric, having tones atElements 1, 5, and 13; and eight rhythms  were  weakly metric,

having tones only at Elements  1 and 13.

There were  19 pairs that contained two identical rhythms,  one

for each of the possible rhythms. To restrict the number of pairs that

contained  two  different  rhythms and to determine if participants

could identify which tone differed between the two rhythms, I

selected the pairs of different rhythms such that the two rhythms of

a pair differed by one tone shifted one grid element. There were 29

such pairs (disregarding the order of the two rhythms).

A  total of 48 pairs of rhythms were used in the  experiment.

Nineteen pairs contained two identical rhythms (all of the possibili-

ties). Twenty-nine pairs contained two different rhythms: From the

possible set of 29 pairs, 21 pairs were presented in one order, and 4

different pairs were presented in both orders (8 in total). Across the

pairs  of  different  rhythms,  the five  instances  in  which  the two

Page 3: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 3/16

1548 HANDEL

rhythms had the same figural organization were included, whereas

the remaining pairs were made up of two rhythms with  different

combinations of metric strength. All individual rhythms occurred at

least once, and nearly all of the rhythms occurred equally  often as

the first or second rhythm of the pair.

Task.  On every trial, two rhythms were presented, and the

participant judged whether the two rhythms were the same ordifferent  in terms of the relative timing of the tones. If the

participants thought that the two rhythms were identical, they

circled the word same on the answer sheet. If they thought the two

rhythms were different, they circled one of five equally spaced Xs

that represented the individual tones in order to indicate which one

had  changed timing  (the  participants were instructed beforehand

that  only one of the three middle tones could change, so that,

effectively, they were to circle one of the three middle Xs). The first

rhythm was always presented with a higher pitch tone (586  Hz,

triangle waveform), and the second rhythm was always presented

with a lower pitch tone (440 Hz, triangle waveform).Presentation rate.  The tones were presented at a moderate rate:

The duration of each grid element was 133 ms, so the length of one

repetition was 2.13 s (16 X 133 ms) and there were 2.3 elements/s.

The interval between Beats 1,5,9, and 13 was about 500 ms, close

to the preferred tempo of  major  beats  (Fraisse,  1982;  Parncutt,

1994). Each tone was composed  of a 10-ms onset ramp, a 50-ms

steady state, and a 10-ms offset ramp (roughly 50% of the duration

of a grid element).

The rhythms were generated  with  BRS-Foringer modules and

were prerecorded and presented to participants on cassette tapes.

The experimental session took place in a small room (3 X 4  m)with acoustical ceiling tile. The participants were seated 2.5 m from

two vertically stacked speakers, each of which presented the entire

rhythm. The rhythms were presented at a comfortable listening

level, approximately 65 dB (SPL); participants were allowed to

adjust the loudness if they wished.

Alternation conditions.  The rhythm-pairs were presented in

two ways. In the first, each rhythm was presented once, and thenthe pair  was recycled three times  (i.e., notated AXAXAXAX to

indicate that the presentation would be AA for pairs of identical

rhythms and AB for pairs of different rhythms). In the second, each

rhythm was presented two times, and then the pair was recycled

(i.e., AAXXAAXX). For both conditions, the  following  rhythm

started at the finish of the three silent grid elements that ended each

rhythm (i.e., at 2.13-s periods). There were no differences between

the repetition of one rhythm and the alternation between  differentrhythms. I chose these two conditions to determine if the number of

repetitions and alternations between the two rhythms  affected

discrimination.

Experimental  design.  The design was within subjects: 2 alter-

nation conditions X 48 rhythm-pairs. Each participant was pres-

ented with two blocks of trials, one block for each alternation

condition. For each alternation condition, three different sequences

of the pairs were constructed that roughly counterbalanced order.

The order of presentation of the alternation conditions and the

sequences within each condition were counterbalanced across

participants.

Procedure.  Two strategies were used to acquaint participants

with  each condition.  First, before the actual presentation of each

alternation condition, there were  four  practice trials that used

simpler four-element rhythms. Two consisted of identical pairs, and

two consisted of different pairs. If participants were confused, then

these rhythms were repeated until participants  felt  confident.Second, the first two rhythm-pairs were repeated later among the

48 experimental trials (thus, there were actually 50 trials per block).

The participants were not told that these were practice trials, and

the results were not used. There was an 8-s interval between each

trial, during which participants made their  responses.  Participants

did not receive any feedback. There was a short break between the

two blocks, and the experimental session lasted about 50 min.

Results

For the  pairs  of  identical rhythms,  the  percentage  of

  identical responses was the only possible  measure. For

the pairs of different  rhythms, there were two measures: (a)

the percentage of responses correctly discriminating the two

rhythms and (b) the percentage  of responses that correctly

identified which tone had changed timing.

Preliminary analyses indicated that there were no  differ-

ences  in  discrimination among  the  three orders  for  each

alternation condition  (the average absolute  difference  was

7%),  that there were  no  differences  between  the first and

second blocks  (79% correct  for  Block  1; 80% correct  for

Block  2), and  that there were  no  differences  between  the

alternation conditions (the average absolute difference across

rhythms was 3%). Thus, the differences  between the control

variables and between the two alternation conditions were so

small  that all the results were combined. The percentage

correct  is  based  on two  responses  from  each  of the 57

participants (114 responses in total), one from each alterna-

tion condition.

Pairs of  identical rhythms.  As described in the  Method

section,  the 19  rhythms  and  resulting pairs  of  identical

rhythms  logically could  be  classified into  four  levels  of

metric  strength on the  basis  of whether tones occurred  at

Element 5, Element 9, both Elements 5 and 9 (5&9), or

neither Element 5 nor Element 9 (none). The results are

shown in Table 1. Overall, there was a significant difference

between  the pairs  of identical rhythms, F(18,  1008 = 3.9,p < .005,  MSB  = 0.11.  The  pairs were then placed into

groups  of  equivalent discrimination performance through

the use of  1\ikey's  honestly significant difference  (HSD)

procedures. These procedures indicated that  the 19 pairs

could be placed into three groups and that each group

contained pairs with  different  metric strength.  The first

group contained (a) the three perfectly metric rhythm-pairs

with tones at both Elements 5 and 9 and (b) Rhythm 16. In

this group, discrimination  was  nearly perfect;  95% of the

responses indicated that the two rhythms were identical. The

second  group contained 10 rhythm-pairs, essentially those

for  which  the percentages correct were  in the 80% range.

The third group contained the five rhythm-pairs with percent-ages correct that were below  80%. These latter rhythms

tended to  have weaker metrics.  In  sum, there  is  some

evidence that metric strength affected discrimination, but the

difficulty  of the  rhythms varied across  a  strict metric

categorization.

Pairs  of  different  rhythms.  These parrs can be classified

in two ways. The first is according to whether or not the two

rhythms had the  same  figural  organization. There were 5

pairs that had the same figural organization and 24 that did

not.  The second  is to create  a 4 X 4  table  defined  by the

metric strength  of the first and  second rhythms (the four

levels being 5, 9, 5&9, or none). For example, one possibil-

ity  could be  defined  by the first rhythm having  a tone  at

Page 4: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 4/16

METRIC AND FIGURAL RHYTHMS 1549

Table 1

Percentages o f  Correct Discriminations fo r  Pairs of Identical Rhythms

for  Experiments  1, 2, and 3

Alternation condition

Rhythm

1.  X...X.X.X...X...2. X...X...X.X.X...3. X.X.X...X...X...

M

4.  X.X...X.X...X...5 X XX X  X...6. X.X..X..X...X...7. X..X..X.X...X...8. X...X.X..X..X...9  X X X  X.X...

10.  X...X..X.X..X...11.  X...X.X...X.X...M

12.  X.X...X..X..X...13.  X.X..X...X..X...14.  X..X...X.X..X...15. X.X...X...X.X...16. X X. X..X..X...17. X..X..X...X.X...18. X..X.X...X..X...19.  X..X...X..X.X...M

Metricstrength

5&95&95 & 9

99995555

NoneNoneNoneNoneNoneNoneNoneNone

Experiment 1 :

AXAXAXAX  Experiment 2:•  AAXXAAXX  AAAAXXXX

Perfectly metric

96959294

Strongly metric or metric

888080868275748681

Weekly metric

807574839286827681

84868685

886460

677675

72

66637178

70

Experiment 3:AAXX

97969797

988992

899293

92

83778491

84

Note.  Empty cells indicate that a rhythm was not used in an experiment.  Indicates whether a tone occurred at Element 5, Element 9, both Elements 5 and 9 (5 & 9), or neitherElement 5 nor Element 9 (None).

Element  5 and the  second rhythm having tones  at  both

Elements 5 and 9 (e.g., X...X..X..X.X...  and X...X...X.X.X...).

In such a table, 5 of the 16 cells cannot occur because of the

restriction that the two rhythms differ only in the position of

one element. For example, this restriction eliminates  a pair

of rhythms in which one rhythm has a tone at Element 5 but

not at Element 9 and the second rhythm has just the reverse.

The pairs are shown in Table 2, those with the same

figural organization appearing first followed by those with

different  figural organizations. Both types  of  pairs wereplaced in tables defined by the metric strength of the first and

second rhythms. The rhythm-pairs with  different  figural

organizations were divided into three groups based on the

relative metric strength of the first and second  rhythms: (a)

The metric strength of the two rhythms was equal; (b) the

metric strength of the second rhythm was stronger; or (c) the

metric strength  of the first  rhythm  was  stronger.  The

percentages of different-rhythm (i.e., correct) judgments and

the percentages of correct element identification are shown.

All of the rhythm-pairs with identical figural organizations

and representative instances of the rhythm-pairs with differ-ent figural organizations are shown (the numbers of pairs are

indicated  in parentheses following  the  instances,  and the

ranges of percentages correct are shown following the

average values). Overall, there was a significant difference

between pairs in terms of the percentage of correct discrimi-

nations, F(28,1568) = 20.0,;? < .005, M SB  = 0.13, and the

percentage  of correct identifications, F(28,  1568)  =  12.1,

p < .005,  MSE  =  0.20. There was no interaction for either

measure.

First consider the same-different figural organization

distinction. There was no overlap in discrimination accuracy

between pairs  of  different  rhythms with  the  same  figural

organization and those with  different  figural organizations

(Tukey's HSD test). For the five pairs with identical figural

organizations, the mean percentage of judgments that the

two rhythms were  different  was 52%, nearly identical to

chance performance, and the mean percentage of correct

element identifications was 23%. Thus, even when partici-

pants did perceive the two rhythms as being different, they

were unable to identify which tone had changed (chance

performance being one third of the trials on which listeners

detected a difference, or 17% in this instance). In contrast,

for the 24 rhythm-pairs with different  figural organizations,

the mean percentage of judgments that the rhythms differed

Page 5: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 5/16

1550 HANDEL

Table 2

Experimen t 1 : Percentages  of  Correct  Discriminations a nd  of  Correct Identifications

of  Which Tone Changed  Position fo r  Pairs o f  Different  Rhythms

Metric strength

First

rhythm

Second

rhythm Rhythm-pan* Discrimination' Identification'Identical  figural  organizations

None

None

5

9

None

5

None

None

X.X..X...X..X...X.X...X..X..X...

X..X...X..X.X...X...X..X..X.X...

X...X..X.X..X...X..X...X.X..X...

X.X..X..X...X...X.X..X...X..X...

X..X.X..X...X...X..X.X...X..X...

M

47

50

51

56

54

52

19

20

22

30

24

23Different  figural organizations

None

5

9

None

5

9

9

5&9

5&9

None

5

9

5

5&9

5&9

None

5

9

X.X...X...X.X...X..X..X...X.X... (6)

X...X..X.X..X...X...X.X..X..X... (3)

X..X.X..X...X...X..X..X.X...X... (6)

M

X..X..X..X..X...X...X.X..X..X... (2)

X...X..X..X.X...X...X...X.X.X...

X..X..X.X...X...X...X.X.X...X... (2)

M

X..X..X.X...X...X..X  X X X

X...X.X.X...X...X...X.X..X..X...

X...X.X.X...X...X..X..X.X...X... (2)

M

83 (70-94)

81 (79-85)

82 (76-91)

83

87 (80-94)

85

87 (80-93)

86

93

94

88 (84-92)

92

51 (37-65)

55 (44-65)

50 (38-61)

52

36 (32-39)

41

42 (32-52)

39

56

60

58 (46-70)

58

'Indicates whether a tone occurred at Element 5, Element 9, both Elements 5 and 9 (5 & 9), or neitherElement 5 nor Element 9 (None). Rhythm-pairs with two possible combinations  of metric strengthsare not shown because of sampling among all the rhythm-pairs. The two combinations are (a) 5followed  by None, and (b) None followed by 9. The numbe r of instances for combinations ofmetric strength  are  shown  in  parentheses following  a  representative pair.  The  percentages  ofcorrect discriminations  and identifications are the averages of all the  possible pairs. The lowest andhighest performances across the rhythm-pairs are shown in parentheses following the averages unlessthere was only one possible instance.

was 81% and the mean percentage of correct tone identifica-

tions was 48% (chance performance is  27%).Second, consider  the metric strength analysis.  What  is

clear is that performance did not differ  among the combina-

tions  of  metric strength  for the two kinds of  different

rhythm-pairs. For the rhythm-pairs  with  identical  figural

organizations,  there were no differences  between pairs in

which both rhythms had a weak metric and pairs in which

the first rhythm had a strong metric (a tone at Element 9).

For the  rhythm-pairs  with  different  figural  organizations,

Page 6: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 6/16

METRIC AND FIGURAL RHYTHMS 1551

there were no  differences  that were due to the grouping

according to relative metric strength that is shown in Table 2,

F 2,21) = 2.8,p > .05. Moreover, there were no differences

that were due to the overall metric strength at the following

four levels: (a) none-none; (b) 5-none and none-9; (c) 5-5

and 9-9; and (d) 5-S&9, 9-S&9, 5&9-S, and S&9-9, F(3,

20) = 0.8. There was no simple pattern of outcomes amongthe  cells.  Consider the two cells that have the greatest

number of rhythm-pairs. Both the average percentage  and

the range of percentages were nearly identical for (a) pairs in

which both rhythms had the  weakest metric structure (no

tones at Elements 5 or 9) and (b) pairs in which both rhythms

had a strong metric structure (tones at Element 9).

There were differences between pairs of rhythms in the

participants' ability to  identify  which tone  shifted  position.

For some pairs of  different  rhythms, the identification of

which tone had changed timing was relatively poor, al-

though it was easy to discriminate the two rhythms. A typical

example was the following pair of rhythms: X..X..X..X..X...

followed  by X...X.X..X..X... In this case, the two rhythmswere judged  as  being  different  in  nearly  all of the  trials

(93%), but the choice of the second tone (underlined) as the

one that had changed was at chance (32%). Instead, partici-

pants usually chose the third tone as the one that had moved

(48%). What is characteristic is that in the first rhythm, there

is an isolated tone (separated by two or more silent intervals

on both sides) that shifts one position to form a double group

in the second rhythm. Participants tend to perceive the

second tone of the double group as having shifted, not the

first tone. This result points out the primacy of the initial

tone of the group; it is perceived as being stable and as

defining the timing of the entire group.

Discussion

In sum, the results give little support to the notion that the

metric strength or metric availability influences the discrimi-

nation between this class of rhythms. If the two rhythms of a

pair had the same figural description, then participants

discriminated the two rhythms only at the chance level,

independent of the metric structure. If the two rhythms had

different figural descriptions, then participants were able to

discriminate them easily, but performance again was indepen-

dent of the metric structure. The only evidence that the

metric structure improved discrimination  was  found  for

pairs of identical rhythms: Rhythms with a perfect metricorganization were more accurately perceived as being

identical.

It is possible to detail the relationship between the metric

and figural organizations  from  two perspectives. The first

perspective involves considering the effect of metric strength

on the five  pairs with identical  figural organizations.  The

rules used to construct the rhythms impose constraints that

limit  the possible rhythms. These constraints yield only a

small set of rhythm-pairs with identical figural organiza-

tions, and within this set, the two rhythms tend to have

weaker metrics. In one  case,  both rhythms have a weak

meter, and in four cases, the rhythms have different strengths.

Performance was  equivalent even though it  might be ex-

pected that a change in meter would highlight differences in

tuning (much the same as would a transition  from tonal to

atonal melodies). This difference in metric strength would be

maximized in the last two pairs, in which the first rhythm

had a strong meter, including a tone at the second strongest

beat at Element 9, and the second rhythm did not. Yet

discrimination was no better than chance.The second perspective involves considering the effect of

figural organization on pairs with identical metric strengths.

In  all instances, there was no overlap in discrimination

between pairs  with identical figural organizations and pairs

with different figural organizations. Consider the pairs with

the weakest metrics:  the none-none pairs.  Discrimination

for  the pair with identical figural organizations was below

chance, whereas discrimination for the six pairs with

different  figural  organizations ranged  from  70% to  94%.

Now consider pairs  with  the stronger metrics:  9-none. In

these pairs, the first rhythm had beats at Elements 1, 9, and

13. Nonetheless, discrimination for the two pairs with

identical figural organizations was barely above chance(54% and 56%). In contrast, discrimination for the pair with

different figural organizations was nearly perfect (93%). To

summarize  this argument, the difference  in  discrimination

between pairs with identical and  different figural organiza-

tions occurs equally for pairs with weaker metrics and for

pairs with stronger metrics. Thus, the predominance of the

figural organization is not limited to weaker metric rhythms.

Even a quick perusal of the outcomes suggests that

discrimination  is  affected  by  many  factors.  For  example,

discrimination involving Rhythm 16,  X..X..X..X..X...,  is

quite good even though the four-beat metric is weak.

Possibly, this rhythm is organized according to a three-beat

metric despite the fact that such a meter would not split therhythm evenly. In this case, the pairs containing Rhythm 16

should be considered highly metric. A reanalysis of the pairs

of  rhythms with  different  figural  organizations  in  which

those pairs containing Rhythm 16 were omitted did not

change any of the outcomes: There was no  differencebetween the groups defined by relative metric strength (in

Table 2) or by overall metric strength as  defined  above.

Thus, the evidence for a three-beat meter is equivocal. The

rhythms might simply be organized according  to the even

spacing of the tones without a meter being induced at all

(without a sense of stronger and weaker beats).

There are several  possible  explanations of why these

results did not show any effect of the metric structure.1. The discrimination task may not have been  sufficiently

sensitive.  The  percentage correct  for  pairs  with  differentfigural organizations ranged  from around 80% to 85%, and

this may have represented  a ceiling  effect  that masked any

effects of the metric structure. Povel and Essens (1985) used

a reproduction task, and Smith and Cuddy (1989) used a

reaction time task; both of these may have allowed meter

effects  to occur. However, even the more sensitive measure

used here, the identification of which tone changed timing,

did not show any differences, and a ceiling effect is unlikely

to have affected that result.

2. The metric structure may not have been  apprehended

strongly. There  are two parts  to  this argument. First, the

Page 7: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 7/16

1552 HANDEL

single or double alternation may have precluded the building

up of the metric structure of either rhythm. These alternation

conditions  were chosen because they produced the best

discrimination in previous work (Handel, 1992), but neither

one  gives  the  listener repeated  looks at the  possible

organizations  to  induce  the  metric organization  of  each

rhythm. Second, these are relatively hard rhythms to pick up.On the  whole, none  of  them  follows  simple rules,  so the

meter may have been relatively hidden. The  rhythms with

the strongest metric structure are the simplest because of the

tone  locations,  so it is  difficult  to  tease  out  clearly  the

contribution of the metric structure.

Taken together, the rapid alternations and the difficulty of

the rhythms might argue that the metric structure did not

affect discrimination because listeners could not pick up and

make  use of the  meter.  To  counter these objections,  in

Experiment 2 I changed the methodology to try to enhance

the perception of the metric structure.

Experiment 2

The goal in Experiment 2 was to investigate  further  the

role of metric structure in the discrimination of rhythms. The

results of Experiment  1 seemed  to show that the  effect  of

metric structure was minimal. Participants organized the

rhythms according to the figural grouping structure, and the

grouping structure, in turn, determined discrimination.  In

Experiment 2, I tried to bring the metric structure into

perceptual prominence  by  making  two  changes  in the

methodology.

The first change was to the alternation condition. In

Experiment 2, the first rhythm was repeated  four times, and

then the  second rhythm  was  repeated  four  times  (i.e.,AAAAXXXX). The rationale for this was that playing each

rhythm four times in a row would allow the listener to pick

up the meter more easily. It is possible that the single and

double alternations used in Experiment 1 forced listeners to

concentrate on the figural structure.

The second change was to introduce a short-duration,

low-pitch pulse tone to  time the first rhythm, in much the

same way that a percussion accompaniment occurs on the

meter beat in a performance. Two approaches were used. In

the first approach, the pulse occurred at Elements 1,5,9, and

13 of the first rhythm whether or not a rhythm tone occurred

at  those  elements. Thus, it was possible to have a pulse

without  a tone. There were no pulse tones for the secondrhythm.  I did not use pulse tones for the second rhythm

because I was afraid that participants would use the derived

strategy of simply attending to the coincidence of pulse and

tone in the two rhythms to determine if they were the same

or  different  (pulse tones were used  for  both rhythms in

Experiment 3). This condition is termed the meter-rhythm

condition because  the pulse timed the underlying meter of

the first rhythm. In the second approach, the pulse occurred

at Elements  1, 5, 9, and 13 only when a tone also  fell  on

those  elements.  The purpose of this condition was to

eliminate  the  possible  confusion  when  a  pulse occurred

without a tone and to emphasize the link between the meter

pulses  and  tones. Because tones always occurred on Ele-

ments 1 and 13, the  difference  between the two conditions

could occur only at Elements 5 and 9. As in the meter-

rhythm  condition,  the  pulse occurred only  on the first

rhythm. This condition is termed the accent-rhythm condi-

tion  because  the  pulse accented  the  tones  at the  metric

positions.

Theoretically, the pulse could improve discriminationequally for strong and weak metric rhythms by providing a

timing reference. Furthermore, this reference would exist

whether or not the tones and pulses coincided. However,

perceptually, it is probably easier  to use  such a reference

when the  pulses  and  tones coincide,  as  they  do for the

stronger metric rhythms, than it is when the pulses and tones

fall  on  different  grid elements,  as they  do for the weaker

metric rhythms. In these latter cases, one might expect the

effect  of the  pulse to be  variable, depending  on the exact

timings of the pulses and tones.

To  summarize,  the  purpose  of  Experiment  2 was to

determine whether repeating each rhythm  four  times and

introducing a pulse would emphasize the metric structuresufficiently  for the strength of the metrical organization  to

affect discrimination.

Method

Participants.  All 61 participants were undergraduate volun-

teers at the University of Tennessee who received course credit for

their participation.  Different participants were used in each experi-

ment. The participants were tested in small groups of from 1 to 3.

Rhythms.  The rhythms were the same type as those used in

Experiment  1. A total of 27 pairs of rhythms were used. Thirteen

pairs contained two identical rhythms. These were selected from

the possible set of 19 (see Experiment 1) so that there was a roughlyequal distribution of metric strength: Three pairs had tones at

Elements 5 and 9, at Element 9, or at Element 5, and  four pairs did

not have tones at either of these metric elements. Fourteen pairs

contained two different rhythms. Five of these pairs contained two

different  rhythms that had the  same  figural  organization,  and the

remaining nine pairs had rhythms with  different combinations of

metric strength. All rhythms were used.

Task.  The task was the same as that used in Experiment 1.

Presentation conditions.  The rhythms were presented at the

same rate that was used in Experiment 1, roughly 2.3 elements/s.

The first rhythm was presented  four  times and then the second

rhythm was presented four times. The timing of the alternation was

the same as that used in Experiment  1. In contrast to Experiment  1,the first rhythm was presented with a lower pitch tone (400 Hz

triangle wave), and the second  rhythm was presented  with a higher

pitch tone (600 Hz, triangle wave). The high and low pitch tones

had   a  steady-state duration of 50 ms and 10-ms/lO-ms rise/fall

times, as in Experiment 1. The pulse was a  100-Hz  sine wave

presented for 35 ms: 15 ms steady state and 10-ms/lO-ms rise/fall

transients. All of the other conditions were identical to those in

Experiment  1.Experimental conditions.  There were three conditions. In the

first, termed  rhythm-rhythm, each rhythm was presented alone

without an accompanying pulse. This condition was identical to the

situation in Experiment 1  (with  the exception of the alternation

method) and served as a replication. In the  second  condition,

termed meter-rhythm, the pulse tone occurred on Elements 1, 5, 9,

and  13 for the  four  repetitions of the first rhythm. If there was a

rhythm tone, the pulse and tone started synchronously, although the

Page 8: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 8/16

METRIC  AN D  FIGURAL RHYTHMS 1553

tone continued  after the pulse. If there was not a rhythm tone, the

pulse  was  heard alone.  In the  third condition, termed  accent-

rhythm, the pulse occurred on Elements  1,5,9, and  13 of the first

rhythm only if a tone occurred. The timing for the pulse and tone

was identical to that in the meter-rhythm condition.

Experimental  design.  The design was within subjects: 3 condi-

tions X 27 patterns. The 27 patterns within a condition werepresented  in a single block, and the blocks were presented to the

participants in the identical  order: rhythm-rhythm, meter-rhythm,

and accent-rhythm. This order was chosen to maximize the  effect

of  the external pulse. The rationale was that the rhythm-rhythm

condition would serve as a reference, allowing participants to

become familiar  with  the rhythms. Following this, participants

could make use of the pulse to perceive the timing structure of the

rhythms without having to become familiar with the rhythms. The

accent-rhythm condition  was  presented last because  it  seemed

somewhat unrepresentative of natural rhythms and I did not want it

to create negative transfer to either the rhythm-rhythm or meter-

rhythm  conditions. For each block, three  different  sequences of

rhythms were constructed that roughly counterbalanced the order

of the pairs of rhythms.

As  in  Experiment  1,  before each block, participants were

presented with four simple examples, and the first two trials in each

block were replicated later within the block. The timing of the trials

was the same as that used in Experiment 1, and there was no

feedback. There were short breaks between the conditions, and the

experimental session took about 50 min.

Results

The  outcomes were analyzed according  to the  samestrategy used  in  Experiment  1.  The  results  for  pairs  ofidentical rhythms are considered first, followed by the

results  for the  pairs  of  different  rhythms. Preliminaryanalyses indicated that performance for the three orders ofeach condition  was  equivalent.  For the  rhythm-rhythmcondition, the percentage correct ranged  from  74% to 71%;

for  the meter-rhythm condition, it ranged from 72% to 70%;and for the accent—rhythm  condition, it ranged from  76% to70%.  On  this basis,  I  combined  the  results  for the  threeorders for each experimental condition. The percentagescorrect are based on 61 responses,  one per  participant.

Pairs  of identical rhythms.  There were  no differencesamong the  three experimental conditions, F(2, 120) =  1.15,

and there was no interaction between rhythm and condition,F(24,  1440) =  1.85. The percentages  of correct judgments

for  the rhythm-rhythm, meter-rhythm, and accent-rhythmconditions w ere 76%, 73%, and 75 %, respectively. Overall,there  was a  significant difference  among  the rhythm-pairs,F(12,  720)  =  15.4, p <  .001,  MSB  =  0.19.  The  bestdiscrimination occurred  for the  three rhythms with  thestrongest metric structure (i.e., with tones at Elements  1, 5,9, and 13) and for  Rhythm  4. For  these rhythms,  thepercentage correct averaged 86% . For the remaining pairs (9in total), there were no differences  as a function of experimen-tal condition or metric strength. Here the average percentagecorrect was  69%.  Thus, the significant  effect  occurredprimarily between rhythm s with the strongest metric and allthe  others.  The percentages  of  correct judgments averaged

across experimental conditions  are shown in Table 1.

Pairs  of  d i f f e r e n t  rhythms.  For all parrs, the percentage

of  judgments that  the two  rhythms were  different  and thepercentage of correct note identifications are shown in Table3 as a function of condition  and of the  metric strength of thefirst and  second rhythms. As in  Experiment  1,  pairs withidentical  and different  figural  organizations were separated.

Overall,  for  both  the  percentage  of judgments that  the tworhythms were different and the percentage of judgments thatcorrectly identified wh ich element had changed timin g, therewere  significant differences between rhythm-pairs, F(13,

780) =  17.5, p < .005, MS B  = 0.21, and F(13, 780) = 9.7,

p < .005, MS B  = 0.22, respectively;  no differences betweenconditions, Fs(2, 120)  = 1.4 and 0.5, respectively;  and asignificant  Rhythm-Pair  X  Condition interaction,  F(26,

1560) = 3.1,p < .005, MSB  = 0.15, andF(26,1560) = 2.7,p < .005, MSB  = 0.15, respectively.

The  Rhythm  X  Condition interaction  is the  crux  of theresults. Consider first the rhythm-rhythm condition, equiva-lent to the  situation  in  Experiment  1.  Here there  was no

overlap in discrimination accuracy between rhythm-pairswith the  same  figural  organization  and  rhythm-pairs withdifferent  figural  organizations (Tukey's  HS D  test). Therewere no  differences among  the  pairs with identical figuralorganizations  and no  differences among  the  parrs withdifferent  figural  organizations. Neither  the  relative  noroverall metric strength  affected  discrimination. For pairswith identical and different figural organizations, the  percent-ages of correct discriminations were 50% and 84%, respec-tively, and the percentages  of  correct identifications of thetone that changed position were  19% and 41%, respectively.

These percentages were nearly identical  to  those  for thesame pairs in Experiment  1.

Consider next the meter-rhythm and accent-rhythm con-ditions. These results  are  quite  different from  those  of therhythm-rhythm condition, and there was a striking interac-tion  between pairs with  the  same  figural organization  andpairs with different figural organizations as a function of themetric strength of the first and second rhythms. For the parrswith the  same  figural organization, there were  tw o  distinctoutcomes.  If the first  rhythm  had the  weakest metricstructure,  with tones  at  Elements  1 and 1 3 but not atElements 5 or 9, then discrimination w as below chance. Thepercentages of correct different judg men ts for the meter-

rhythm  and  accent-rhythm conditions were 46% and 40%,respectively (this difference was not  significant).  However,

if the first rhythm had a stronger metric structure, with a toneat Elements 5 or 9, then the pulse in the meter-rhythm andaccent-rhythm conditions generated  significantly better per-formance. The percentages of correct different judgm entsfor  the  meter-rhythm  and  accent-rhythm conditions were72%  and 69%, respectively.

For the pairs with different figural organizations, with twoexceptions, discrimination was excellent. The percentage ofcorrect different judgmen ts was 76%,  and the percentageof  correct element identifications averaged  39%. As wasfound  for  the, rhythm-rhythm condition, there were  noeffects  that were due to metric strength. Even excluding thetwo  pairs with the poorest discrimination, the percentage

correct was slightly h igher for the rhythm-rhythm condition.

Page 9: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 9/16

1554 HANDEL

Table 3

Experiment 2: Percentages  of  Correct Discriminations Discr.) and of Correct

Identifications Ident.)  of Which Element Changed Position for Pairs

of   Different  Rhythms

Experiment condition

Metric strength*

Firstrhythm

Secondrhythm Rhythm-pair

Rhythm-rhythm

Discr. Ident.

Meter-rhythm

Discr. Ident.

Accent-rhythm

Discr. Ident.

Identical figural organization

None

None

5

9

None

5

None

None

X.X..X...X..X...

X.X...X..X..X...

X..X...X..X.X...X...X..X..X.X...

M

X...X..X.X..X...

X..X...X.X..X...

X.X..X..X...X...X.X..X...X..X...

X..X.X..X...X...

X..X.X...X..X...

M

44

49

47

43

57

57

52

14

20

17

18

23

19

20

53

38

46

75

75

67

72

16

18

17

39

31

31

34

36

44

40

61

67

80

69

14

20

17

25

25

30

27

Different figural organization

None

5

9

None

5

5

5&9

None

5

9

5

5&9

None

9

X..X..X...X.X...

X..X..X..X..X...

X.X...X..X..X...X.X...X...X.X...

X...X.X..X..X...X...X.X...X.X...

X..X..X.X...X...X..X.X..X...X...

X.X...X.X...X...X.X...X..X...X...

M

X..X..X..X..X...X...X.X..X..X...

X...X..X..X.X...X...X...X.X.X...

M

X...X.X...X.X...X..X..X...X.X...

X...X.X.X...X...

X..X..X.X...X...

M

85

74

82

87

92

84

90

87

88

77

84

81

 Indicates whether a tone occurred at Element 5,Element 5 nor Element 9 (None).

43

33

39

41

48

41

48

23

36

44

54

49

Element 9,

85

52

62

84

82

73

82

79

81

74

84

79

43

20

31

43

38

35

36

36

36

34

55

45

84

53

54

83

82

72

80

84

82

79

92

86

55

26

23

46

48

39

26

33

30

35

61

48

both Elements 5 and 9 (5 & 9), or neither

_ _  „*  < • _ _  _«i *___ _ _ ?̂ *t.

Thus, the pulse did not improve discrimination if the two

rhythms had different figural organizations.

There were no differences  among the three conditions in

the participants' ability to identify  which tone had changed

position. More important, the

 percentages of

 correct identifi-

strength in Table 3. There are large differences  among the

pairs. But,  in general, participants were unable  to  identify

more accurately  that a tone had moved from a strong to a

weak metric position than the reverse.

Page 10: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 10/16

METRIC AND FIGURAL RHYTHMS 1555

Discussion

These results, completely consistent with those from

Experiment  1,  clarify  the  role  of the  metric structure  in

rhythm perception. The meter did improve discrimination

for  pairs  of  identical rhythms,  but the  effect  was  more

complicated  for  pairs  of  different  rhythms. To  follow  theform of the Discussion in Experiment  1, let us first consider

the effect of metric strength on the five pairs of rhythms with

the same figural organization. For the rhythm-rhythm pairs,

performance hovered about chance even if the initial rhythm

had tones at the strongest metric beats (at Elements 1 and 9).

Thus, these results perfectly replicate those of Experiment 1.

For the meter-rhythm and accent-rhythm pairs, the external

pulse brought about a fixed temporal grid that enabled

listeners to distinguish between two rhythms with the same

grouping or figural organization but  with  different  timings

between the groups. However, the pulse improved discrimi-

nation only if the pulse coincided with the tones of the initial

rhythm. The pulse made the metric structure of the  initialrhythm more available so that the timings between  adjacent

groups could  be  encoded relative  to the meter. However,

when the pulse did not coincide with any of the three internal

tones of the first rhythm (the first two pairs in Table 3), the

pulse did not affect discrimination even if the second rhythm

had a stronger meter.

Now consider the effect  of the figural organization on the

pairs with the same metric strength. For the rhythm-rhythm

pairs, the results replicate those in Experiment 1. For all

comparisons, discrimination  for pairs with  different  figural

organizations  was  well above chance,  in  contrast  to the

chance performance  for all  pairs  with  identical  figural

organizations.  For the  meter-rhythm  and  accent-rhythmpairs, the results  differed  because there was overlap among

pairs with different and identical figural organizations.

This asymmetry, in which a transition  from  stronger to

weaker structure produces better performance than does  a

transition from weaker to stronger structure, has been  foundin  other kinds of auditory processing. For example, Jones

and  Boltz (1989) found that a hierarchic  (i.e.,  stronger

meter) to nonhierarchic (i.e., weaker meter) transition yielded

better performance than the opposite sequence.  Bharucha

and Pryor  (1986)  found  a similar asymmetry in temporal

discrimination. Krumhansl, Bharucha, and Castellano (1982)

showed that harmonic discrimination  was better if the first

melody  was strongly tonal and the  second atonal than thereverse, and Bartlett  (1993)  demonstrated that  listeners

could detect changes in melodic contour more easily if the

tonal contour preceded the atonal contour.

Jones and Boltz (1989) argued that all of these kinds of

results can be subsumed under the concept of expectancy.

The initial melody, rhythm, or sequence of tones creates a

trajectory  about subsequent events. Stronger meters (or

more tonal melodies) more tightly constrain  the  range of

expectancies so that rhythmic or tonality deviations can be

more easily perceived (expectancy is roughly analogous to

Garner's, 1974, notion of inferred sets). Nonetheless, it

seems reasonable that the structure of the second sequence,

by  inducing a set of expectancies or possible alternatives,

ought to affect discrimination. But that was not the case here.

There are two possible reasons: (a) The second rhythm was

being directly compared with the first rhythm without being

encoded,  or (b) the  fact  that there  was no  pulse  for the

second rhythm made the metric strength less prominent for

the listener.

Experiment 3

What is known at this point is  that the  metric strength

influences  discrimination  in  only  one  context:  if the two

different  rhythms have the same figural organization, if the

first rhythm has the stronger inherent meter, and if the metric

elements of the first rhythm are marked by a pulse. I

designed Experiment 3 to investigate two issues  further.The first issue is why the effects of the metric structure are

so limited.  One possibility is that because the pulse never

occurred for the second rhythm, the effect of the pulse in the

first rhythm was weakened  and any possible  effect  of the

metric strength of the second  rhythm was negated. For thisreason,  in one condition  in Experiment  3 I had the pulse

occur on elements  1, 5, 9, and 13 for both rhythms (termed

the meter-meter condition) in order to continue the metric

structure across the second rhythm. These outcomes can be

compared with those when the pulse occurred for the first

rhythm only (meter-rhythm)  or not at all (rhythm-rhythm).

(Although participants usually could determine if the two

rhythms  were  different  for the meter-meter  condition  by

simply determining if the coincidences of pulses and tones

were identical in the two rhythms, no participant explicitly

reported using this derived strategy).

The second issue concerns the generality of the asymme-

try in outcomes that is due to the relative metric strengths ofthe  two rhythms when there is a pulse. I argued for this

asymmetry in Experiment 2 by comparing across rhythm-

pairs. For example, discrimination was at chance for Pair A,

in which the meter was weak-strong, but discrimination was

very good for Pair B, in which the meter was  strong-weak.

Because any pair of rhythms have unique timing characteris-

tics, it is impossible to argue unambiguously that a strong-to-

weak  meter transition always yields better discrimination.

For this reason, each pair of different rhythms with the same

figural organization  was  presented twice,  the  second  time

with the order of presentation of the two rhythms reversed. If

there is a general asymmetry created by the interaction of

relative metric strength  and  pulse, then discrimination  of

each combination of rhythms should be better if the rhythm

with the stronger metric is presented first than if the rhythm

with the weaker metric is presented first. In pairs in which

both rhythms have equal metric strength, there should be

little difference between the two orders of the rhythms.

Method

Participants.  All 67 participants were undergraduates  at the

University of  Tennessee  who  received course credit  for  their

participation. They were tested in groups of from 1 to 3.

Rhythms.  The rhythms were  the same type  as  those used  in

Experiments  1 and 2.  There were  36  pairs  of  rhythms  in  all.

Thirteen pairs had two identical rhythms; these were the same pairs

Page 11: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 11/16

1556 HANDEL

used  in Experiment 2. Twenty-three pairs had two  different

rhythms. These can be broken into two sets. The first set consisted

of the five pairs of different  rhythms used previously that had the

same  figural  organization. Each pair  was  presented  in the two

possible orders to generate a total of 10 pairs. The second set

consisted of 13 pairs that had different figural organizations. These

pairs were chosen to match the metric strengths of the pairs with theidentical figural organizations.

Task.  The task was simpler than the one used in Experiments 1

and  2. In  Experiment  3  participants judged whether  the two

rhythms  were identical or  different  on a 4-point scale (1 =  very

sure identical,  2 = fairly sure iden tical,  3 = fairly sure  different,

and 4 =  very  sure different).  In deriving the percentage  correct, I

considered judgments of 1 and 2 to be  identical judgments and 3

and 4 to be different judgments. I switched the response in order

to  maximize performance so that any possible improvement that

was due to the  meter could  be  found, and I  felt that trying  to

identify the tone that changed position could be distracting.

Presentation conditions.  The rhythms were presented at the

same rate used previously, roughly 2.3 elements/s. The first rhythm

wa s presented two times, and then, without a break, the second

rhythm  was presented two times. This alternation condition waschosen to make the task more difficult so that any improvement that

was due to the meter would not be obscured by a ceiling effect. The

rhythms were generated with the MIDI software package MIDI-

LAB for the IBM PC. The sounds were generated with a Seiko

DS-250 keyboard, and the piano timbre was used. The first rhythm

was  presented with a lower pitch tone (440 Hz), and the second

rhythm was presented with a higher pitch tone (660 Hz). Each tone

was 75 ms in duration. The pulse was a 99-Hz tone presented for

40ms.

Experimental design.  The design was within subjects: 3 condi-

tions  X  36 pairs of rhythms. The rhythms were presented in three

blocks according to condition and across subjects; the order of

presentation of the conditions was counterbalanced. For each

condition,  the  rhythms were  placed  into  four  sequences  thatcounterbalanced order. The procedures were the same ones used in

Experiments 1 and 2. There were short breaks between the three

conditions, and the experimental session took about 75 min.

Results

Preliminary results indicated that there were  no  differ-ences that were  due to the  sequence order within each

condition. The maximum difference  in the percentage cor-

rect between  the  sequences  for any  condition  was  11%.

Moreover, there were no  differences  in percentage correct

that  were due to the order of the blocks. The  percentages

correct for  Blocks  1, 2, and 3 were 77%, 82%, and 82%,

respectively. On this basis,  the results were averaged oversequences and order. The percentages correct are based on

67 responses, one per participant.

Pairs of identical rhythms.  There was a small significant

difference  among conditions, F 2,  132)  =  3.9, p <  .025,

MSB   =  0.09,  a significant Rhythm X Condition interaction,

F(24, 1584) =  2.6, p <  .005,  MSB  = 0.07, and a  significantdifference among rhythm pairs, F(12, 792) = 9.8, p < .005,

M SB   = 0.08. For the three rhythms in which tones occurred

at all four meter elements, discrimination was nearly perfect.

The  percentages correct  for the  rhythm-rhythm,  meter-

rhythm,  and meter-meter  conditions were 99%, 95%, and

96%, respectively. Discrimination was equivalent for the six

rhythms in which tones occurred at Elements 1,5, and 13 or

at Elements 1,9, and 13. (The only exception was Rhythm 4:

The  percentage correct equaled that for rhythms  with  the

strongest meter). The percentages correct for the rhythm-

rhythm,  meter-rhythm, and  meter-meter  conditions were

89%,  91%,  and  96%, respectively. Discrimination  was

poorer for the four rhythms that had tones only at Elements 1

and  13. The percentages correct for the rhythm-rhythm,

meter-rhythm, and meter-meter conditions were 88%, 78%,

and  84%, respectively. Over all pairs, the differences among

conditions were small, averaging 2%. For this reason, the

percentage correct  for each rhythm was  combined across

conditions and is shown in Table  1.

On  the  whole, discrimination  for the  pairs  of  identical

rhythms mirrors that in the first two experiments, although

the percentage correct is higher in Experiment 3. What is

common across the three experiments is that it was easier to

identify  two strongly metrical rhythms as  being identical.

What  differs  in Experiment 3 from  Experiments  1 and 2 is

the hint of an interaction between metric strength and pulse

condition. Discrimination was identical across the various

pairs of identical rhythms for the rhythm-rhythm condition

without  any  external pulse,  but  discrimination  was  more

difficult  (i.e., participants were more likely to hear the two

rhythms as being different) for the less metrical rhythms for

conditions with an external pulse.

Pairs of  different  rhythms.  Fo r these pairs, there were asignificant  rhythm effect,  F(22,  1452)  =  33.0, p <  .005,

MSB   = 0.22, a condition effect, F(2, 132) = 4.1, p <  .025,

M SB   =  0.21, and a Rhythm X Condition interaction, F(44,

2904)  =  4.3,  p <  .005,  MSB  =  0.15.  The  Rhythm X

Condition interaction is the main focus of the results (as in

Experiment 2). This interaction can be understood from

three perspectives.First, consider the overall difference between the 10 pairs

with  identical  figural  organizations  and the 13 pairs with

different figural  organizations. As  found  in  Experiments  1

and 2,  there were large  differences  in the  percentage

of  correct discriminations  for the  rhythm-rhythm  and

meter-rhythm conditions,  F(l, 21) =  58.0, p <  .001, and

F(l, 21) = 6.5, p < .02, respectively. Discrimination was at

the chance level for pairs with identical figural organizations

(52%)  and well above chance for pairs with different figural

organizations (77%). These percentages are nearly identical

to those  found  in Experiment 2. In contrast, there was no

difference  in discrimination for the meter-meter condition

because discrimination  for the parrs with identical  figuralorganizations improved sharply (63%) compared with dis-

crimination for  pairs with  different  figural  organizations

(77%),F(1,21) = 2.5 /».05.

Second, consider  the 10 pairs  of  different  rhythms with

the same figural organizations. For the two pairs in which the

tw o rhythms had equal metric strengths (la and Ib in Table

4) and the  four pairs in which the rhythm with the weaker

metric preceded the rhythm with the stronger metric (2a-

5a),  discrimination was at the chance level or lower and was

essentially equal for the three conditions.  The only excep-

tions  occurred  for  Pairs Ib  and 5a for the  meter-meter

condition.  In  contrast,  if the  stronger metric rhythm pre-

ceded  the weaker one  (2b-5b),  discrimination was much

Page 12: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 12/16

METRIC AND FIGURAL RHYTHMS

Table 4

Experiment 3:  Percentages o f  Correct Discriminations fo r  Pairs  of Different  Rhythms

1557

Metric strength

Firstrhythm

Second  Pairrhythm no.

Experimental condition

Rhythm-pair  Rhythm-rhythm  Meter-rhythm Meter-meter

Identical figural organization

None

None

None

5

9

None

None

None

5

9

None  (la)

(Ib)

5  (2a)

(3a)

9 (4a)

(5a)

None  (2b)

(3b)

None  (4b)

(5b)

None (6)

(7)

5 (11)

(12a)

9  (13a)

(14a)

None  (12b)

None (13b)

(14b)

X.X..X...X..X...X.X...X..X..X...

X.X...X..X..X...

X.X..X...X..X...

M

X..X...X.X..X...X...X..X.X..X...

X..X...X..X.X...X...X..X..X.X...

X.X..X...X..X...

X.X..X..X...X...

X..X.X...X..X...X..X.X..X...X...

M

X...X..X.X..X...

X..X...X.X..X...

X...X..X..X.X...X..X...X..X.X...

X.X..X..X...X...

X.X..X...X..X...

X..X.X..X...X...X..X.X...X..X...

MDifferent  figural

X.X...X..X..X...

X.X...X...X.X...

X..X..X...X.X...X..X..X..X..X...

X. X X X X .X...X.X..X..X...

X..X..X...X.X...X...X.X...X.X...

X.X...X..X..X...X.X...X.X...X...

X X X X XX..X..X.X...X...

M

X...X.X...X.X...X..X..X...X.X...

X.X...X.X...X...

X.X...X..X..X...

X..X..X.X...X...X  X X X  X...

M

41

50

46

45

33

38

36

38

56

57

53

65

58organization

73

93

91

85

86

91

88

74

91

96

87

32

63

48

19

29

41

40

33

70

75

77

75

74

47

85

72

77

77

88

80

73

93

95

87

49

72

61

32

26

62

65

46

80

69

92

77

80

45

69

91

84

83

82

85

87

96

97

93

(table  continues)

Page 13: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 13/16

1558 HANDEL

Table 4  (continued)

Metric strength3

First  Secondrhythm  rhythm

Pairno.

Experimental condition

Rhythm-pair  Rhythm-rhythm Meter-rhythm Meter-meter

Different  figural  organization

5

9

5  8a)

 8b)

9 9)

 10)

X...X.X..X..X...X...X.X...X.X...

 

X...X.X..X..X...

X.X...X.X...X...

X.X..X..X...X...

X..X..X.X...X...X..X.X..X...X...

 

58

7

8

77

76

42

6

62

70

6

66

71

65

60

  Indicates whether a tone occurred at Element 5, Element 9, both Elements 5 and 9 (5 & 9), or neither

Element 5 nor Element 9 (None).

better and there were large differences between the condi-

tions.  For the  rhythm-rhythm condition, discrimination

improved  from 38% to 58%  although performance  was

above chance only for Pair  5b. For the meter-rhythm  and

meter-meter conditions,  the improvement  was much more

dramatic  (from  33% to 74% and  from  46% to 80%,respectively)  and discrimination of all four pairs was above

chance. These results are shown in Table 4.

Th e discrimination performance  for the  four pairs going

from a weaker to a stronger metric breaks into two classes.

Performance was better  for the meter-rhythm condition an dconsiderably better for the meter-meter condition for Pairs

4a and 5a  than  for Pairs  2a and 3a.  Probably,  in the twoeasier pairs,  the  tone that changed moved  to  Element  9,

which created a stronger metric than in the two other cases in

which the tone moved to FJement 5, which created a weaker

metric. There was no such difference for the rhythm-rhythm

condition,  which demonstrates the potential usefulness of

the meter if the figural organization does not discriminate

two  different rhythms. The difference  in discrimination between

the meter-rhythm and meter-meter conditions for Pairs 4a

and  5a illustrates that a pulse that accompanies the stronger

meter of the second rhythm can improve discrimination.

Third, consider  the 13 pairs in which the two rhythms haddifferent figural organizations. For the rhythm-rhythm con-

dition, discrimination for all pairs (with the exception of Pair8a) was equivalent. There was no effect  of metric strength.

Moreover, as described above, discrimination  for all of these

pairs was better than that for pairs with the identical figural

organization.  For the meter-rhythm  and meter-meter condi-

tions, discrimination was far more variable, and the effect of

metric strength was inconsistent. Discrimination was poor-

est when the two rhythms in a pair had the identical metric

strength even if both rhythms had strong meters (Pairs 9 and10). Discrimination improved when  the two  rhythms h addifferent  metric strengths  and was  maximum when  thestronger metric  was the first  rhythm  of the  pair.  Tw orhythm-pairs were significantly more  difficult  than  theothers (Pairs  6 and 8a), and performance on them equaled

that for pairs with identical figural organizations. These two

pairs were the same ones that produced poorer discrimina-

tion  in Experiment 2.The identical pattern of results was found even for the best

25% of the participants. These participants, although they

performed  more accurately across all the pairs, also were

unable to discriminate Pairs  la-5a  for the rhythm-rhythm

and  meter-rhythm conditions  and  Pairs  la-3a  for themeter-meter condition.

Discussion

These results support the contention that for these rhythms

and tasks, the metric structure operates within the figural orgrouping structure.  If the figural  organizations  of tworhythms are different, then adding an external pulse will no timprove and may even degrade discrimination. However,

the  metric structure  can  help listeners discriminate  two

different  rhythms with identical  figural  organizations.  Butthe metric structure cannot operate  backward in time.

Unless  the timing of the tones of the initial rhythm fits themeter (i.e., tones occur synchronously with the pulse at beat

elements), discrimination will  not  improve.  In  fact,  if thetones of the initial rhythm do not fit the meter, the pulse will

not act as a temporal grid and will make discrimination more

difficult. Thus the effect of an external meter pulse is context

dependent.

General Discussion

The results of these three experiments help to delimit the

role of the metric structure. The notion of a meter underlying

the perception  of every rhythm, or that metric rhythms areprototypical, i s seductive, bu t that notion may overstate theimportance of a meter in discrimination.

For the rhythms used in these three experiments, listeners

appear  to initially place  the tones into groups based on theshortest interval,  and  those groups  form  the  basis  of thefigural  organization.  If two  rhythms have  different  figuralorganizations, either i n terms of the number of tones in each

group or in terms of the order among the groups (e.g., 1-3-2as opposed to 3-1-2), then listeners easily perceive that the

Page 14: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 14/16

METRIC AND FIGURAL RHYTHMS 1559

two rhythms are different. If the figural organizations are the

same, then listeners must use the timing between the onsets

of  the  initial elements  of  each group  to  discriminate  the

rhythms. It is here that the meter can provide a grid to time

the intervals between successive groups. What is  surprising

is the relative difficulty  listeners have in making use of the

possible grids. When these rhythms were presented withouta pulse, there was little evidence that the occurrence of tones

at the metric elements affected discrimination. The strength

of the inherent meter produced by the timing of the tones did

not  create  a  grid that allowed participants  to  compare

timings across groups. When these rhythms were presented

with a pulse, the pulse improved discrimination  only if the

initial rhythm was strongly metric so that the pulse occurred

synchronously with tones. Otherwise, the pulse did not

improve discrimination and actually impaired it, even though

it  should logically have instituted  the  same underlying

metric and timing relationships. In sum, in certain instances

the meter can be used to distinguish among rhythms with

identical figural organizations by creating a temporal grid, inthe same way that melodic tonality can be used to distin-

guish among equivalent melodic contours by creating a

harmonic template or that facial features can be used to

distinguish among equivalent profiles by creating expressive

prototypes.

Generality  of Results

Comparisons across  experiments.  One  relevant issue

concerns whether the results reported here can be attributed

to  general experimental design  effects,  response biases, or

both. For example, Yee, Holleran, and Jones (1994) reported

that performance for highly skilled participants changed as a

function  of the types of rhythms included within an experi-

mental session. It is  unlikely that such design  effects  could

explain these results even though the three experiments used

a variety of presentation conditions, including  different

response modes, pulse conditions, and alternation condi-

tions, and  used  a  different  set of  participants  in  each

experiment. I argue this on two grounds. First, performance

was remarkably consistent across experiments, particularly

for the critical  different pairs with identical figural organiza-

tions. Consider the five pairs that occurred in all three

experiments, which are the first five pairs shown in Table 2.

The range in the percentage of correct discriminations for

these pairs for the rhythm-rhythm condition across the three

experiments is 6%, 5%, 13%, 4%, and 11%. The range in the

percentage of correct discriminations for these pairs for the

meter-rhythm condition between Experiments  2 and 3 is

21%, 19%, 5%, 2%, and 8%. Second, the use of alternative

pulse conditions (meter-rhythm and  meter-meter)  led to

new response patterns for those conditions without changing

the responses to the other conditions within Experiments 2

and  3. The meter-rhythm condition (in Experiment 2)

yielded better discrimination for strong-to-weak meter pairs

but did not change the discrimination for the rhythm-rhythm

condition. The  meter-meter  condition (in Experiment 3)

yielded better discrimination  for pairs in which the second

rhythm had a tone at Element 9 (Pairs 4a and 5a in Table 3)

but did not affect discrimination for the rhythm-rhythm and

meter-rhythm conditions.

It is also unlikely that response biases  affected  the

outcomes. In each of the three experiments there was a

higher percentage of different pairs (60%, 52%, and 64% for

Experiments 1,2, and 3, respectively) than of identical pairs.

Thus, purely on the basis of frequency, participants shouldhave  responded different if  they were uncertain.  But

participants judged  different  pairs with identical figural

organizations  as  identical, going against  the  probability

expectancy. I believe that participants' initial hypothesis in

this task is that the two rhythms are identical and that they

judge them as  different  only when they perceive  a timing

change. Thus, when the participants do not pick up the

variation in the  between-groups  timing  for two  different

rhythms with identical figural organizations, they judge the

rhythms as being the same.

Alternative  meters.  Participants may be using alterna-

tive meters in perceiving  these rhythms. In that case,  the

analyses based on four-beat meters would not be relevant formany  pairs. Rhythm  16 (in  Table  1) could  be  organized

according  to a  three-beat meter because  the tones  fall  on

Elements 1, 4, 7, 10, and 13. Similarly, Rhythms 7 and 17

begin with tones on Elements 1,4, and 7. In all three of these

rhythms, however, the interval between the onset of the final

tone at Element  13 and the  onset of the next repetition  at

Element 1 is  four  elements, so a strict three-beat meter

would progressively  fall  out of phase with the rhythmic

tones across repetitions. This  difference  in timing is quite

noticeable (399 ms vs. 532 ms). Nonetheless, to avoid the

possibility that the classification of Rhythm 16 as weakly

metric biased  the results, I redid all of the analyses of pairs

of  different  rhythms in Experiments 2 and 3 and omittedpairs involving Rhythm 16. The results did not change for

any  analysis. However, there is some evidence, particularly

in  Experiment 3, that pairs of  different  rhythms involving

Rhythm 16 were easier to discriminate. Overall, the percent-

age of correct discriminations for pairs involving Rhythm 16

was 88%, whereas the percentage of correct discriminations

for  the other  different  pairs with different  figural organiza-

tions  averaged 72%. What makes the conclusion that

participants were using  a three-beat  meter comparison

problematic is that Rhythm 16 has a unique figural organiza-

tion  of  five  individual  elements and  differs from  all of the

other rhythms, which have at least one group of two or more

elements. For this reason, the discrimination could be madewithout any perception of meter at all.

Does feedback  f f e c t  discrimination?  It is possible that

discrimination would have been better, particularly for pairs

of different  rhythms with the same figural  organization, if

participants had been given feedback. This feedback would

have  allowed the participants to learn to attend to the

relevant  timing information between groups. Although I

could not  provide feedback  for  each trial because  of the

available equipment, I attempted to test this possibility with

a fourth experiment in which I provided intensive pretrain-

ing  on the rhythms. Before participants began  a block of

trials  for one  condition,  four different  kinds  of  pairs  of

rhythms  were demonstrated  and  explained:  (a)  pairs  of

identical rhythms that participants misjudged rarely;  (b)

Page 15: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 15/16

1560 HANDEL

pairs of identical rhythms that participants misjudged some-

times (discrimination usually was good for identical rhythms);

(c) pairs of  different  rhythms that participants misjudged

rarely;  and (d)  pairs  of  different  rhythms with  the  same

figural organization that participants judged correctly at less

than  chance. For each kind of rhythm, several examples

were  presented  auditorily while participants  looked  at avisual representation of the rhythms indicating which ele-

ment had shifted (similar to the representations  shown in the

four tables). After the participants stated that they could hear

that  the rhythms were identical or  different  and could

explain the visual representations, the experimental trials

were started. To keep the experiment simple, I used only the

rhythm-rhythm and meter-meter conditions. The orders of

the rhythm-pairs and conditions were counterbalanced,  and

the presentation conditions were identical to those in Experi-

ment 3. A total of 40 undergraduate students participated.

In spite of the extensive pretraining, discrimination was

identical to that in Experiment 3. The most important results

concern the pairs of different rhythms with the same figuralorganizations. For the meter-meter condition, if the rhythm

with the stronger meter preceded the one with the weaker

meter, the percentage  of correct discriminations  was 87%,

but the percentage of correct discriminations for the reversed

pair was  only 40%. (The analogous percentages  from

Experiment 3 were 80% and 46%). For the rhythm-rhythm

condition, if the rhythm with the stronger meter preceded the

one  with  the  weaker meter,  the  percentage  of  correct

discriminations  was  64%,  but the  percentage  of  correct

discriminations for the reversed pair of rhythms was 40%.

(The analogous percentages  from  Experiment 3 were 58%

and  38%). Thus, although  the  pretraining  may  have  im-

proved discrimination slightly for the  stronger-to-weakermetric pairs, it did not improve performance for the more

difficult weaker-to-stronger metric pairs. Overall, the pattern

of results was identical to that in Experiment 3. Specifically,

the pulse improved discrimination only if the stronger metric

rhythm was the first (i.e., presumably the reference) rhythm.

Otherwise, discrimination for the meter-meter and rhythm-

rhythm  conditions  was  equivalent  and  remained below

chance for rhythms with the identical figural organization.

Is  the figural organization  always predominant?  In this

context,  the figural  organization determined whether  two

rhythms  were perceived as identical or  different.  These

results are  contrary  to the  belief that the  meter underlies

rhythmic perception  and that the figural organization repre-sents a weaker or less organized percept. There are several

possible reasons for the present outcome. First, the partici-

pants were unselected and probably relatively untrained, and

metric organization improves  with  musical training (e.g.,

Yee et al., 1994). Second, the discrimination task itself might

have  minimized  the  usefulness  of the  meter,  and the

alternation between rhythms might have made  the meter

difficult  to  perceive. Third,  the pairs  of  different  rhythms

with  identical figural organizations tend to have weaker

meters. Rhythms with stronger meters might have allowed

participants to discriminate between two  different rhythms

with  identical figural organizations. Fourth, the intervals

between the onsets of adjacent tones are based on only oneunit  of 133  ms,  and the  ratios between  different  onset

intervals were 1:2:3, which do not readily fit any simple

meter.

Any one or any combination of these possibilities could

account for the weak effect of metric strength. Nonetheless, I

do not believe that they undercut the  conclusion that the

figural organization  is  primary. A  meter  can  resolve  the

temporal  relationships  among  the figural  groups,  but themeter could not exist without the figural groups. The meter is

an emergent property of the auditory grouping in the same

way  that symmetry  is an  emergent property  of  visual

grouping.

Is a Concept  of  Meter Necessary?

The concept of a meter is deeply ingrained in music

theory and empirical research. Nearly every theorist argues

for the multilevel nature of a regular beat and the interplay

between the metric and grouping structure (Yeston,  1976).

Moreover, the most influential recent treatment of the

subject, by Lerdahl and Jackendoff  (1983), argued that the

metric and grouping structures are independent.

Similarly, there are extensive findings in both motor and

music research that there is a basic temporal unit that is used

to create  all of the  intervals.  Fraisse (1982)  summarized

research demonstrating that  when  participants reproduce

intervals between three or  four  tones, the intervals are

systematically distorted so that they approach one of two

values: a short interval that is roughly one half the length of a

long  interval.  Povel  (1981)  and  Povel  and  Essens (1985)

argued that listeners attempt to create an internal clock (i.e.,

a meter) when listening to a rhythm. Rhythms that are easily

encoded  are  those  in  which accented tones  fall  at  equal

intervals, and listeners are assumed to search for a clock that  hits the accented tones. In addition, Jones and coworkers

(Jones, 1993) have argued for a model of rhythmic attending

in which the presence of a regular pattern of strong and weak

beats allows the listener to extrapolate the passage and

identify  the important structural elements. Finally, Collier

and Wright (1995)  suggested that there are innate prefer-

ences  for specific rhythmic ratios  and that musicians can

scale these ratios faster or slower. But even experienced

musicians find it difficult to learn complex ratios, and none

could scale these complex ratios. What this all means is mat

there is a consensus that rhythmic production is based on a

temporal unit (or possibly  two independent units) and that

this unit is subdivided  or multiplied to produce the variousintervals  in a  passage.  Using this unit  as a  reference,

performers can stretch or contract durations to produce the

desired artistic effect  (Clarke, 1985).

The problem, then, is to reconcile the overwhelming

evidence for the importance of a meter in reproduction tasks

to the relatively small effects found for a meter in the present

discrimination tasks. One way is to make use of ideas  frominformation  theory that Garner (1974) used to explain

differences  in recall and discrimination  tasks. All theorists

have explicitly or implicitly argued that the meter simplifies

rhythms by making them temporally predictable (and thus,

in Garner's terms, making the  inferred set  smaller). Thus,

there  is a simple contingency between  the beats and thetones that fall on the beats. The interval between strong beats

Page 16: The Interplay Between Metric and Figural Rhythmic Organization

8/12/2019 The Interplay Between Metric and Figural Rhythmic Organization

http://slidepdf.com/reader/full/the-interplay-between-metric-and-figural-rhythmic-organization 16/16

METRIC AND FIGURAL RHYTHMS 1561

is always the same, and this allows the listener to extrapolate

the beat and tonal rhythm into ttie future (cf. Jones & Yee,

1993). This regularity allows the person producing a rhythm

to preplan a repetitive motor response for the strong beats

and  to  attend  to the  more irregular intermediate weaker

beats. The memory load is reduced, and therefore we would

expect that reproduction of metric rhythms would be moreaccurate than reproduction of nonmetric rhythms. However,

the simple contingency among the stronger beats of metric

rhythms makes them less differentiated, and therefore we

would expect discrimination among metric and nonmetric

rhythms to be equivalent because of two competing factors.

On  the one  hand,  it is  easier  to  encode metric rhythms

because  of the  simple  contingencies,  and that should  im-

prove discrimination. On the other hand, the simple contin-

gencies make  the  metric rhythms more similar,  and that

should make discrimination more difficult. In these experi-

ments, the pulse improved discrimination, presumably by

allowing participants  to more  efficiently  encode  the metric

rhythm  so that it could be better differentiated  from  thefollowing nonmetric rhythm. This argument points out once

again  the contextual nature of rhythmic organization: The

usefulness of any organization depends on the specifics of

the task, and different concepts may be necessary as the task

varies.

From this perspective, the better reproduction of metric

rhythms  occurs  not  because they are  metric, but  because

they  are  simpler,  and  simpler would  be the  preferred

explanation because  it is applicable  in a broader domain.

There is an inevitable confound here in that metric rhythms

are  necessarily simpler because  the strong beats occur  at

equal temporal intervals  and are  therefore more  easily

predicted. I do not believe  that this  is merely  a semanticissue because it  goes  to the heart of whether rhythms are

simply one type of pattern. If this is the case, then concepts

applicable  to a broad cross-section  of visual and auditory

perceptual phenomena, such as regularity, symmetry, good

continuation, common fate, and similarity, might suffice  to

explain  rhythm perception without the  need  for rhythm-

specific  concepts.

Although I have argued previously that both visual and

auditory events are perceived within a hierarchical  space-time framework  and that  it is  possible  to  form  many

equivalences for these sorts of concepts (Handel, 1988), the

ubiquitous theme-plus-variation structure of music tends to

make the concept of periodicity (i.e., a metric beat) synony-

mous with repeatability, and that need not be true for the

visual  world. We do need  a concept  of  meter  (if  only  to

emphasize the motoric and affective components of rhythm),

but the effect of the metric regularity can be understood only

within the context of the listening task, and  different  tasks

will require different  concepts.

References

Bamberger, J.  (1978).  Intuitive  and  formal  musical knowing:Parables of cognitive dissonance. In S. S. Madeja (Ed.), The arts,cognition  and  basic  skills  (pp. 173-209).  St  Louis,  MO:CEMREL.

Bartlett, J. C. (1993). Tonal structure of melodies. In T. J. Tighe &

W .  J. Dowling  (Eds.),  Psychology  and  music  (pp.  39-65).Hillsdale, NJ: Erlbaum.

Bharucha,  J. J., &  Pryor,  J. H.  (1986). Disrupting the isochronyunderlying rhythm: An asymmetry in discrimination.  Perception&  Psychophysics,  40 , 137-141.

Bolton, T. L. (1894). Rhythm. American Journal  o f Psychology,  6,145-238.

Clarke,  E. F.  (1985).  Some aspects  of rhythm and expression  inErik Satie's Gnossienne No. 5. Music Perception, 2, 299-328.

ColUer, G. L., & W right, C. E. (1995). Temporal rescaling of simpleand complex ratios in rhythmic tapping. Journal of  ExperimentalPsychology: Human Perception a nd Performance,  21 , 607-627.

Essens,  P.  J.,  &  Povel,  D. J.  (1985).  Metrical and nonmetricalrepresentations of temporal patterns. Perception  &  Psychophys-ics, 37, 1-7.

Fraisse,  P. (1982). Rhythm  and tempo. In D.  Deutsch (Ed.), Thepsychology   o f music (pp. 149-180). N ew York: Academic Press.

Garner, W. R. (1974). T he processing of  information  and structure.Potomac, MD: E rlbaum.

Handel, S.  (1974). Perceiving  melodic  and  rhythmic  patterns.Journal of Experimental Psychology,  103, 922-933.

Handel,  S.  (1988).  Space is to time as vision is to audition:Seductive but misleading. Journal  of  Experimental Psychology:Human Perception and Performance,  14 , 315-317.

Handel,  S.  (1992).  The  differentiation  of  rhythmic structure.Perception  & Psychophysics,  52, 497-507.

Jones,  M. R.  (1993).  Dynamics of  musical  patterns:  How domelody and rhythm fit together? In T. J. Tighe & W . J. Dowling(Eds.),  Psychology  and  music  (pp. 67-92).  Hillsdale,  NJ:Erlbaum.

Jones,  M. R., &  Boltz,  M.  (1989).  Dynamic  attending  andresponses to time. Psychological Review, 96, 459-491.

Jones, M. R., & Yee, W . (1993). Attending to auditory events: Therole of temporal  organization.  In S.  Me Adams & E.  Bigand(Eds.),  Thinking  in  sound  (pp. 69-112).  Oxford,  England:Oxford U niversity Press.

Krumhansl, C . L .,  Bharucha, J . J., & Castellano, M . (1982). Keydistance  effects  on  perceived harmonic  structure in music.Perception  & Psychophysics,  31, 75-85.

Lerdahl, P., & Jackendoff, R. (1983). A generative  theory  of  tonalmusic. Cambridge, MA: M IT Press.

Palmer, C., & Krumhansl, C. L. (1990). Mental representations formusical meter.  Journal  of Experimental Psychology:  HumanPerception and Performance,  16, 728-741.

Parncutt, R.  (1994).  A  perceptual model  of  pulse  salience  andmetrical  accent  in  musical  rhythms.  Music Perception,  11,409-464.

Povel,  D.-J.  (1981).  Internal  representation of simple temporalpatterns. Journal  of  Experimental Psychology:  Human  Percep-tion and  Performance,  7, 3-18.

Povel, D.-J., & Essens, P. (1985). Perception of temporal patterns.Music Perception, 3 , 411-440.

Smith, K.  C.,  &  Cuddy,  L. L.  (1989).  Effects  of  metric  andharmonic rhythm on the detection of pitch alterations in melodicsequences. Journal  o f Experimental Psychology: Human Percep-tion and Performance,  15,457-471.

Yee, W ., Holleran, S., & Jones, M. R. (1994). Sensitivity to eventtiming  in regular and irregular sequences: Influences  of musicalskill. Perception  & Psychophysics,  56, 461-471.

Yeston,  M.  (1976).  Th e  stratification  of musical rhythm.  NewHaven, CT: Yale U niversity Press.

Received August 20,1996

Revision received August 8, 1997

Accepted September 15 1997  •