Download as pdf or txt
Download as pdf or txt
You are on page 1of 21

The Construction of Musical Syntax (I)

Author(s): Benjamin Boretz


Source: Perspectives of New Music, Vol. 9, No. 1 (Autumn - Winter, 1970), pp. 23-42
Published by: Perspectives of New Music
Stable URL: http://www.jstor.org/stable/832192 .
Accessed: 25/06/2014 10:37

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Perspectives of New Music is collaborating with JSTOR to digitize, preserve and extend access to Perspectives
of New Music.

http://www.jstor.org

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
THE CONSTRUCTION OF
MUSICAL SYNTAX (I)*
BENJAMIN BORETZ

1. THE NOTION OF REFERENCE


IN THE absence of lexical or conventionalreferentsforthe ordering
of musical elementsin a "semanticallyinterpreted"structure,the con-
structionof coherent musical entitiesdepends on a contextuallyin-
ferrablereferentialbasis. Such dependence requires that it be possi-
ble to distinguishamong individual musical things,and to distinguish
classes of such things,on no more elaborate a priorbasis than thaton
which those things are just identified as musical things in the first
place. If our subject were (verbal) language, the claim thatthe above
requirement could be met would amount to a claim that given only
notionsof what "language" was, of the kind of data thatwas pertinent
to the constructionof "linguisticentities,"and of the varietiesof iden-
tityand functiondesignable forlinguisticentities,itwould be possible
to inferfromthe perceptualquality-patterns of successionsof sounds
and/orinscriptionsassociated withany "linguisticutterance"the par-
ticularvocabularyand syntaxthroughwhichto interpretthe "mean-
ing" and "structure"of thatutterance- i.e., to distinguish itas a particu-
lar linguisticindividual.Thus "English,"whileitwould be a cognitively
designable stage in the definientialascension from"language" to "this
utterance,"would no longer be an essentialstage to account forbythe
introductionof new assumptionsand conventions,since all the essen-
tialcharacteristics of "thisutterance"would be determinablethrough
constructionon the original primitiveswithout the mediate inter-
ventionof a conventionalEnglish dictionaryand grammar.
This designabilitybut non-essentiality of a crucial linguistic"stage"
is, of course, unavailable in verballanguage, at least presently,but the
analogous stagesof determinationin music,i.e., of "tonal,""12-tone,"
etc. determination,do present such a methodological superfluity;
they are convenient but not essentialstages to distinguishexplicitlyin
the constructionof individual musical entities. And thus musical
structuresdo seem ratherspecial among "linguistic-type" entities(see

* From Meta-Variations,Part III. (Parts I and II have


appeared in PNM,Vol. 8, No. 1
and Vol. 8, No. 2, respectively.)
* 23'

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
PERSPECTIVES OF NEW MUSIC

Part I, PNM, Vol. 8, No. 1, pp. 58ff.);fortheirentire"meaning" or, to


be less fanciful,theirsignificative relations,and the determinationsof
complexityand coherence that measure their "cognitivecontents,"
can be derivedsolelyfromtheconceptuallyguided processingof such
"musicallybasic" informationas the perceived patternsof difference
among the qualia of individual auditoryconcreta in associated suc-
cessions,and of degreesand kindsof differenceamong sub-complexes
of such concreta inferredfrom("sliced out of") the total succession.
The degree to which a musical structureapproaches maximalco-
herence (i.e., maximal individuality)is proportionateto the degree to
whichthe metricallydeterminatedata are multiplyexhausted bysuch
slicings,as membersof sets of interrelated"relation-class"extensions.
Such sets,in most "western"music,are understood as tracing"paths
throughthe data" whose totalitypresumablyconstitutes"the musical
structure."
The kinds of informationprocessed in the pitch-timedimensions,
and the fundamentalrelationsin termsof whichthe processingtakes
place conceptually-perceptually, are the subjectsof the second partof
thisessay. The presentsectionbegins at the point where the general
("universal musical") resourcesof relation are marshaled into inter-
connected sets of particularrelationalsyntactical functions.Through
such functions,ultimately,maximallycoherentindividual structures
may be constructed (or reconstructed)as particular orderings-in
particularsequences and multiplicities-ofparticularvalues of these
functionsin particular(quantized) empiricalinterpretations. There is,
thus,compositional"latitude"up to thislast,"notational"stage-and
beyond that there is latitudein the "articulative"decisions of "inter-
pretative performance" which have to do with determiningwhat
happens in perceptualdimensionswhose mostminutedistinguishable
aspects of identity,similitude-differentiation, or proportionalityare
not considered to be syntactically determinateor determining.
The room for "choice" at every level of the construction,both in
concrete interpretationand in more formal determinations-even
well below the first"syntactical"stages-seems to evidence thatmaxi-
mum determinacyin every dimension of structureis a prerequisite
for maximum "creativefreedom,"if thatis considered to be signifi-
cative onlywhen whatis "created" is "something."For the "freedom"
to create and perceive vacuously is hardlyto be preferredto a com-
pletely conventionalized "determinacy,"where the "rules" are not
chosenbut given(in both compositionand audition); in both cases the
"musical individual"is severelycircumscribedin the degree to which
it can be perceived as "individual" or is even "identifiable"at all, by
preciselythe degree to whichitsperceivablesare determinablyinter-
. 24'

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
THE CONSTRUCTION OF MUSICAL SYNTAX (I)

pretable, in the firstcase, and contextuallyinterpretable,in the


second. For to the extentto which a "lexical-conventional"predeter-
mination is brought to musical structures,they are merely "repre-
sentative"("in a language"), and, as such, non-particular.Hence, no
"ideas" emerge except above the highest such lexical-conventional
level. But withoutreferencesfortheconsistentmeasurementof simil-
itudes and differentiaat all, no "characteristics"can emerge either,
except in the most general dimensions (such as the dimensionsthat
seem to be principallyfunctional in some recent music, by, e.g.,
Penderecki and Xenakis) where the metricizationsare restrictedto
aspects of sonic perception not conceptually unique to music, as
(grossly)"louder-softer,""denser-thinner,"etc., dimensionswhichin
maximallydeterminatemusic are merelypart of the articulativesur-
face through whichthe structurallyrelevantpaths throughpitchcom-
plexes are indicated.
Thus, since our music is largely pitch-structural music, the most
favorablecentralreferentsformostof itsinstancesare single,particu-
lar partitioningsof the pitch domain. The "singleness"of such par-
titioningsis, of course,fundamentalto our notionof "totalstructure,"
a notion whichis expressed implicitlyany timewe speak of a piece as
"the piece" or anythingthatcharacterizesit as "a thing."Such parti-
tionings,in existingmusical systems,make possible a hierarchization
of relationsand theirrelationswhich,in the most fertileof such sys-
tems,also makes it possible to overlap the pitchdomain manifoldlyin
termsof determinaterelationsat hierarchically distinct
levels(generating
the "distinctmeanings" for the "same relations"referredto above).
Thus a musical syntaxis essentiallya model forthe determinationof
the interlocking structure of hierarchically connected relations
through which the range of significationsof a discriminableset of
data can be interpreted.The more "efficient"the model, the larger
the number of distinctand unambiguously determinate relational
functionsthat can be inferredfroma single datum-succession.This
standard of unambiguousness, however, is often misunderstood;
withoutdwellingon the point,I suggestconsiderationof the essential
differencebetween unambiguousmultivalence in which an event has
multiple "meanings," simultaneouslycotenable and perceivable but
not "identical,"and ambiguity, where it is not determinablewhether
any one particularrelational"meaning,"of a given nature,is signifi-
cantly ("meaningfully,"determinably) ascribable to the event in
question.
And since there are two principal bases for the constructionof a
partitioningreferentialset,and itturnsout thattheseare mutuallyin-
compatible,it is just here that our systemexperiences its firstmajor
* 25'

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
PERSPECTIVES OF NEW MUSIC

"fork."The imperativesthat underlie thisdivergenceare the subject


of the sequel.

2. CONTENT-CENTRICITY AND ORDER-DETERMINACY


Given that a presented pitchcollectionmay be perceptuallyidenti-
fied in terms of three factors,dimension("number of token-slots"),
content("identityand relation of distinctelement-types[i.e., things
representablebydistinctprimitivesymbols]representedbythe totality
of element-tokens"),and the orderof the content-tokens,it is evi-
dentlypossible to resta pitch-syntactical systemon the discriminative
basis of eitherof the lattertwo(the first,dimension,alone could func-
tion as a primaryreferentonly in a non-pitch-basedmusical syntax
since it cannot be used to take into account pitchidentitiesand rela-
tions,and thereforethe identitiesof the partitioningsets withinany
sound-presentationalarray would have to be determined by other
means than theirpitchaspects). Consider, forexample, the pitchsuc-
cession (however interpreted)(A-A-Bb-A). If tokendimensionalone
were determinate,we would have as the identityof this setjust the
quantity4, as the "number of slots" in the set:
(Xo, xl, X2, X3)

If, however,we definecontent


to withinjust pitchidentity/nonidentity,
we obtain the result:
(x, x, y, x)
by which we can assert distinct contentdimension (2, as (x,y)); orderof
content representation(x-y-x); dimensionof primitive-symbol multi-
plicity(3x, ly); and finally,order of distributedtokens (x-x-y-x).
If, further,the token-referentsare intervallicallyquantized,it is
furtherpossible to assertrelationsamong the elements,as:
1. content= (0, 1) (where x = 0, y = 1)
2. order of content= (0 1 0)
3. content-tokenorder = (O 0 1 0)
As far as interpretation
is concerned:
1. At the dimension-determinate level (A-A-Bb-A) is equivalent to
any four-token pitch set.
2. At the content-dimension-determinate level, any sets of whatever
dimensionor order whose tokensare all tokensof one of exactlytwo
distinctpitchelements,such thatthereis at leastone tokenin each set
of each pitch element,are equivalent. Thus (A-A-Bb-A)= (C-F-F-C-
C-F-C). (2a.) With token-dimensiondeterminacyadded to content-
dimension determinacy(A-A-Bb-A)= (C-C-C-F).
26

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
THE CONSTRUCTION OF MUSICAL SYNTAX (I)

3. With content-identity
determinacy,any two sets representingjust
two distinctpitch elementsrelated by the interval1, each of the ele-
ments of which may be mapped into an element of the other by the
identityoperation,are equivalent.Where the lattermapping is notby
the identityoperation but by the addition of or subtractionfrom a
given integer, such two sets are transpositionallyor inversionally
equivalent,respectively.Thus, in the firstcase (A-A-Bb-A)= (Bb-Bb-
A-Bb-Bb-A-A); and in the second (or the third) case (A-A-Bb-A)=
(C-B-B-C-B). (3a.) Where token-dimensiondeterminacyis added to
content-identity determinacy(A-A-Bb-A)= (Bb-A-A-A).
4. At the order-of-content-determinate level, any two sets of any
token-dimensionsin which the firsttokens are tokens of the same
pitch element,and the next distinctpitch elements representedare
the same in both sets,etc.,are equivalent.Thus (A-A-Bb-A)= (A-Bb-
Bb-A-A). (4a.) Withtoken-dimensiondeterminacyadded to order-of-
content determinacy(A-A-Bb-A) = (A-A-Bb-A). Transpositionally
and inversionallyrelated sets are equivalent under this condition,
also.
Thus we may speak of content-determinate pitch systems(or con-
tent-referential, or content-generative,systems),and order-deter-
minatepitchsystems,notingthatwhereorder is syntactically determi-
nate this involves a more complicated referentialbasis than where
contentalone is, since "order" is not independent, but a referential
orderingof the contentof a pitch set.
But what do we mean by "syntacticalsystem"or "syntacticalfunc-
tion"?And how are "content"and "order" inferredas "bases" of musi-
cal structures?Briefly,this is what is meant: if a particular "slice"
made of auditory experience consists of an ordered array of pitch
events,a "musicalstructure"may be inferredas theglobal interpreta-
tion of thisarrayby means of a series of more local partitionsof the
array into proper subsetsof pitches(or even intojust a singlesubset
equivalent to the whole, but that is in one respecta trivialand in an-
other respect a nonstandard instance). These partitioningsare not
necessarilylimitedby temporaladjacency; and all partitioningsof the
actual pitchdata are conceived in termsof some inferredpartitioning
of the pitch or the pitch-classdomain which functionsas a unitary
referent(in dimensionand content,or dimension,content,and order)
foreverypartitioningsubsetof the presentedarray.Referenceto this
referentis enabled throughinferenceof a delimitedset of mapping
functions(e.g., "transposition")bymeans of whichthe "isomorphism"
of the array-partitioningsets to the referentialset and to one another
can be asserted. The sequences in whichsuch mapping functionsare
applied, and the rationale through which their successive applica-
* 27 -

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
PERSPECTIVES OF NEW MUSIC

tions are interrelated,maximallyspecifyingthe characteristicsof the


presentedarrayin termsof the referentialaspect(dimension,content,
or order) of the referentialcollection
(thus maximallydeterminingthe
partitioningsubsets,or, rather,producing a maximumhierarchically
ordered set of subset-typesin whichadjacent typesare minimallydis-
tinct,whose least member is as close to having a monadic contentas
possible, and in termsof each of whichthe entirepresented succession
can be ordered), can be considered the "structure"of thecomposition;
and whatever is chosen as the "background" systemof hierarchies,
functions,and relations may be considered the syntax, and the hier-
archies,functionsand relationsthemselvesmaybe consideredsyntacti-
cal ones.

Retrospectively,then,the definitionsof Part II can be regarded as a


hierarchizedspecificationof basic relationsand functionsbymeans of
which the references for any pitch-syntactical systemcan be con-
structed.We have, in turn, specifiedcontentoperations and order
operations and shown them to be conceptuallyand formallyanalo-
gous (that is, we have seen that "transposition,""complementation,"
and "partition"are analogouslyoperativeon bothcontentand order).
But here we need to considerwhatgeneralconceptscan be explicated
as the cognitivedimensionsin termsof whichobservablecontent-and
order-differentiafunction to produce references for syntactical
functions.
When we talkof "partitioningsubsets"and theirinterlockinginter-
relationsas "defined" by referential(unordered or ordered) sets of
primitiveelementsand "rules of relation,"whatwe are talkingabout
is comparison: comparison of such subsetsto othersof the same sub-
set-typelevel,in termsof theirconstituentproper subsets(of,in turn,
the same [subset-of-subset]-type levels), as well as in termsof their
relativetime-orderpositions.1Thus when we speak of "content-deter-
minate"systemswe mean systemssuch thattheprimary basis of subset
comparison determinedby them (that is, the basis forcomparisonof
subsetswhose content-[or token-]dimension is equivalent to thatof
the principal referenceset,whichthusbegins withcomparison at the
most "background" structurallevel but is not limited to it) is the
measurementof contentintersection,both of itsdegree(amount),and
1 What constitute
boundaryconditionsformembershipin a givensubset-typecannot
be stipulatedindependentlyof contextualdetermination, just as the referentialaspect
of the referentialset is not uniquelydeterminablefroman examinationof the contents
of the set itself,alone. Even the relevantdimension-limitation
is notdeterminableinde-
pendentlyof context.Thus, for example, at some levels of some compositionin some
syntacticalsystem(C,D,C) and (C,D) may be membersof the same subset-type, whileat
some other level of the same composition(not to mentionsome other compositionin
the same syntaxor in a differentsyntax)theymay not be membersof the same subset-
type.
28 -

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
THE CONSTRUCTION OF MUSICAL SYNTAX (I)

its nature(relationalcontent).These two factorsare then interpreted


as the determinantsof the degree of "P"-similitudeexhibitedbycom-
pared subsets, for some "P" such that "P" is a relevant"syntactical
function"withinS. (There may be several independent,or interde-
pendent, functionalpredicatessubstitutablefor"P," as, e.g., "interval
similitude,""pitch-elementsimilitude,""token-dimension-per-pitch-
element similitude,"etc.)
Now to determinedegree of simple pitch-elementsimilitudeonly
identity,number, and positionof discreteelement-typeswithinsub-
sets to be compared need be defined,as:
A = {ao, a, . . ., ai}
B = {bo, b, . . ., bj}
C = {co, C1, ... , Ck}
D = {do, dl, . . . , dl}
For any pitch-structure S, where A is the referenceset for S, and B
and C are proper subsetsof S, if B and C are primary subsetsof S (at
their respectivestructurallevels), then all the tokensof B as well as
all those of C representdistinctelement-typesand i =j = k, or (in a
relativelyweaker system)thereare in A, B, and C thesame numberof
distinctelements. (If token-dimensionalone were stipulated as the
relevant"syntacticalcriterion,"its capacityto functionas such would
depend just on the satisfactionof the conditionthatthe primitiveele-
mentswere all of the same general type,e.g., "all pitches,"etc.) Now
for a structureto be content-determinate, the subsets regarded as
discrete withinit must be comparable (and differentiable in varying
degrees and ways)on the basis of theirelement-typecontent.Thus if
we know only that
bo = w Co= x
bl=x and cl=y
b2= y c2 = z
b3 = z C3= t
and j = 3 and k = 3 (i.e., B = {w, x, y, z} and C = {x, y, z, t})2 then,
without knowing anythingabout the quantized interpretationsof
w, x, y,z, t (except forthe conditionw # x # y # z = t,or, w, x, y,z, t
are distinct),we can content-comparethe sets as follows:
B U C= {w, x, y, z, t}
B C = {x, y, z}

2
I.e., dimension of B = dimension of C; thus,while the "referenceset" definesjust
the referential-aspect dimension of compared subsetsof "primarydegree," here that
referential-aspectdimension (i.e., distinctelement-typedimension) is, in fact,identical
with the token-dimensions of the membersof this particulararray of "primary-degree"
subsets.
29

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
PERSPECTIVES OF NEW MUSIC

B-C ={w}
C-B ={t}
B+C ={w, t}
and, of course, $ (B) and $(C) will have the same dimension and:
(B) = {{ } {w}, {x}, {
, {},},, }w,z} {w, , {w } w z, , { }, x, z},
, { , x, wy
{y, z} , { , x, z} , {w, , z}, {w,, x,
y, , y,, z}}
$(C) = {{ }, {x}, {y}, {z}, {t}, {x, y}, {x, z}, {x, t}, {y, z}, {y, t},
{z, t, {x, y,{x,
z},
z {x, t}, x , {, y, z, t}, {x, y, z, t}}
, t}
x, y,
and:
83(B) n (C) = {{ } , {y},{z, {x, y}, {x, z}, {y, z}, {x, y, z}}
{y},
$(B) - (C) = {{w}, {, , {, , ,, {w, z},z, {w,w, x, y},x,{w,
w, x, z},
{w, y, z}, {w, x, y, z}}
$(C) - $(B)= {{t}, {x, t}, {y, t}z,{z, t}, {x, y, t}, {x, z, t}, {y, z, t},
{x, y, z, t}}
3(B) + (C) = {{t}, {w}, {w, x}, {w, y}, {w, z}, {x, t}, {y, t}, {z, t},
{w, x, y}, {w, x, z}, {w, y, z}, {x, y, t}, {x, z, t}, {y, z, t}, {w, x, y, z},
{x, y, z, t}}.
Similar observationscould be made regardingthe number and con-
tent of all possible subsetsof S of the given reference-aspectdimen-
sion (since the reference dimension is regarded as definingthe refer-
ence-aspect dimension of a maximum relevantsubset of S; all larger
subsetsare "described"as multiplerepresentationsof such a subsetor
"compositions"of several such subsets,as: {w, x, y,z, t} maybe "un-
derstood" as B U C, etc.,but notas a potentiallyprimary"syntactical"
subset in itself),withinthe total set S (the totalpitcharray). S itselfis
thus alwaysdescribableas one instanceof a successionof instancesof
the referenceset (i.e.,of definedtransformations of the referenceset);
each such description occupies a "structurallevel," and a conse-
quentially ordered succession (or "nest") of such descriptionscon-
stitutes"the structure"of the S in question.
If the element typesare quantizationallyinterpretedas, e.g., pitch
classes, then the referenceset may specifynot only token-and con-
tent-dimensionbut also content-relation; and operations may be de-
fined,forany two subsetsof a givensyntactically relevantsubset-type,
to determinetheircontent-relationalisomorphism.3'4
3These operations, like every theoreticalterm defined on a set of primitivesin a
system,serve to reduce the number of distinctvocabularyelementsby assertingfunc-
tional correlationsfor classes of such elements. Thus every sound-tokenin an array
must be taken as a distinctpitch element before the assertionof the propertyof pitch
whichreduces the numberof distinctelementtypesbyitsapplication.Similarly,
identity,
before intervalidentityis asserted,everydistinctpitchpair is a "different
interval"(since
intervalis just "the relationdeterminedby two pitches").Again, reductionin vocabu-
lary(distinctintervaltypes)and increasein functionality (multiplepossible [distinguish-
* 30 -

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
THE CONSTRUCTION OF MUSICAL SYNTAX (I)

Thus, in our present example, we can describe not only the degree
(amount) of pitchintersection, but the identity(pitchcontent)thereof,
though not its relational nature (intervalcontent):
if B - C = {w}
and C-B = {t}
and D = {v,w,x,y}
so that B - D = {z}
and D- B = {v}
and C - D = {z,t}
and D - C = {v,w}
then we can assert a hierarchyof "degrees of closeness" and one of
"kindsof closeness"among B, C, and D. If A is the referenceset,and
A is {u,v,w,x},then we can also assert"closeness to A," definedas the
inverseof "differencefromA." So, if E is a presented subsetin S, and
E = {u,v,x,y, then
A-E={ }
and A - E.
In this case, we can "map" E into A by an identity operation. But we
have no way of mapping B, C, or D into A or into each other; forour
specificationtellsus only of the identity-nature
of the intersection,not
for {u,v,w,x,y,z,t}
its relation-nature; could be anypitches.
Thus,
if u = 0, v = 3, w= 4, x = 6, y = 8, z = 9, t = 1,

able] interpretations of each type,and functionalindependence, foreach pitchpair, of


its pitch-vocabulary"meaning"and itsinterval-vocabulary "meaning") result.Further,
intervalvocabularyis again reduced by interval-classcorrelation(mod-n-"octave'-
equivalence) and then by complementary-interval-class ("inversional")correlation,and
pitch vocabulary is reduced by pitch-classformation,generated out of interval-class
correlation.Thus, the operationsof transpositionand complementationon complexes
of pitchesand intervalssimilarlyreduce the "distinctset type' vocabularyand increase
the functionalrange of each set type (i.e., enable us to regard distinct
setsas distinct
in-
stancesofsinglesettypes);and theydo so byextendingthe same principlesof intervaland
interval-classidentity (for transposition)and interval-class-complement correlation
(for inversion)throughwhichthe basic vocabularyof musicalelementswas reduced by
the elaboration of the basic vocabularyof musical functionsat the mostgeneral levels
of theirdefinition(see Part II, PNM, Vol. 8, No. 2).
4 In the tonal is the only necessarysuch operation, but comple-
system,transposition
mentation is also significative
at everylevel, even though everycomplementationof the
diatonic collectionis describableas a transpositionthereof(T7I of a "major" collection
produces the "tonic minor" T3S, while the "dominant"and "subdominant"collections
are produced by the "complementarytranspositions"T7S and T5S, respectively.The
T7I-TOS relation(particularlyin viewof the interval(7) of complementationinvolved)
is especiallysignificantin our laterviewof the "major"-and "minor"-orientedgenera-
tions of the diatonic collectionas inversionallysymmetrical).
31

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
PERSPECTIVES OF NEW MUSIC

then A = {0,3,4,6}
B= {4,6,8,9}
C= {6,8,9,1}
D= {3,4,6,8}
E = {0,3,4,6}
all the relations described above stillhold, even though there is no
transpositionthatwillmap any twononidenticalsetsintoone another.
When complementationis added, B = TO I (D),5 but no similarrela-
tion exists among any other pair, even though they can clearlybe
ranked in "degree"-of-intersection
hierarchywithrespecttoA:

A= 0,3,4,6}
1. E = {0,3,4,6} ("0 different")
2. D = {3,4,6,8} ("1 different")
3. B = {4,6,8,9} ("2 different")
4. C = {6,8,9,1} ("3 different")
And each could be a "referenceset" fora similararrangementof the
others; also the "identities could order an arrayeven
of the differences"
where more than one set had an equivalentamountof differencefrom
the reference:
if F ={ 11,0,3,4} ("1 different")
G = {10,11,0,3} ("2 different")
H= {7,10,11,0} ("3 different")
then the "arraywithrespect to A" would look as follows:
H G F E D B C6

A
for F-- D = {11,O}
and F- = 6,8} ("2 elementsdifferent")
while D-C= (3,4)
and C - D = {9,1} ("2 elementsdifferent")
5 TOI
(D) = "The transposition-zeroset of the mod-n complementset of D."
6 Here the "identitiesof differencefromA" are associated withrelativeamounts of
intersectionamong the sets being compared to A; hence the "linearity"of the array.
If, for example, there were a set I whose contentswere {2,0,3,4}, it would create a
new "path," which mightor mightnot 'join" one of the other paths at some point:

FF <-- E--,
E D
(A)
32

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
THE CONSTRUCTION OF MUSICAL SYNTAX (I)

so that"numberof places away fromN along the array"is equivalent


to "number of elementsdifferentfromN"; in particular,"number of
places away fromE in the array"is an index of "number of elements
differentfromA (the referenceset)". But direction, or relativedirection
away from E along the array yields no additional information about
"contentrelationto A," but only about the content-intersection rela-
tions among the sets along each "directionalvector."
But ifTt (x) = (x + t) (transposition)and Ir (x) = (r - x) (complemen-
tation) such thatx is a pitch-classelement number and t ranges over
the domain definedby the totalityof pitch-classelementnumbers,are
admitted as syntactical("f-") operations, then only D and B, and G
and H have assertiblef-relationsto one another; but there existsno
possible referenceset such that bothpairs have the relation to it. In
other words,thereis no t such thatTt or Ir + Tt willmap D or B into
G or H (where r = 12); forif therewere any such t,therewould be at
least one possible referenceset K such thatD, B, G, and H could all be
mapped into K by various Tt or Ir+ Tt operations,and the relations
of D, B, G, and H to one another could be characterizedin termsof
such a relationto a common reference.
In such comparabilityof all primarysubsets of an S by means of
such a "relationto a common reference"reststhe centricity of the ref-
erence set that is inferred for a content-determinatesystem.The
degreeofcontent determinacy defined by such a systemdepends on the
extentto which the referenceset is (or can be) regarded as being iso-
morphically partitioned by, successively,subsets of decreasing dimen-
sion, where the subsetsof the referenceset determinedby such par-
titioningsare also regarded as "syntactical"-i.e.,normativeforevery
instanceof the referenceset at everylevel of the structure.7The ref-
erence set, thus, may be regarded as generated bythe conjunctionof
sets at the immediatelypreceding subset-typelevel. This latter"or-
dering" is essential if all subsets so arrived at are to be regarded as
"syntactically significant."
Thus the minimum degree of content determinacyis collection cen-
or determinacyof contentof the referencecollectiontakenas a
tricity,
whole. This is the necessaryand sufficient condition forprimary con-
tentdeterminacy of a syntax.Next are the various degrees of construct
i.e., the inferencesof complex subsets (constructsthat are
centricity,
7 Hence
they would define subsettypesrelevantto everystructurallevel, though not
accountingfor all subset typesappearing at all structurallevels (except in the unlikely
case where the "syntax"and the "structure"were totallyindistinguishable,a situation
which is logicallypossible [as a territoryis a map of itself]but practicallyundesirable
since it would representas a "general relation"everythingassertibleabout a particular
piece, withnothingin that piece designated as constitutinga particularinterpretation
chosen froma range of possibilitiesconstrainedby the general relation).
* 33 -

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
PERSPECTIVES OF NEW MUSIC

regarded as "normatively"[i.e., for all of a piece] generatingor par-


titioningthe reference collection isomorphically;e.g., the triadas a
three-elementgenerator of the diatonic collection,the "fifth"as a
two-elementpartitioningthereofwhose own (4 +3) partitioninggen-
erates the triad,then the octave itselfas partitionedby the fifth,and
then the unit pitch domain as partitionedby the referentialoctave).8
So we have, in order of weaker to strongerdeterminacy:

(determiningthe membershipof an element or a


collectioncentricity
relationin the referencecollection; the range of membershipsof an
element or a relation in the member collectionsin the "syntactical
array" determined by the reference set; and the range of "within-
collections" and "between-collections"relationshipsdetermined by
any pair or larger complex of elements);
construct (orderingthe elementsof the referencecollectionin
centricity
termsof "withina construct"and "betweenconstructs"relations,and
possibly [but not necessarily]hierarchizingthe constructswith re-
spect to a "central" constructregarded as generating "subsidiary"
constructsthrough defined transformations, such that the union of
centraland subsidiaryconstructsgenerates the referenceset. At this
8 Thus the
"pitchcentricity"of the "tonal system"seems to be an importantaspect of
it even presystematically(despite, e.g., Randall's attempt,now withdrawn,to generate
it noncentrically),one that seems stronglyinferrablefromthe Schenkerianmodel, for
example. Although the diatonic collection (even as triadicallypartitioned) does not
itself "predetermine"or "presuppose" pitch or constructcentricity,nor, obviously,
need it be generated by an initiallypitch-centricor construct-centric approach, nor,
equally, need it functiononlyin a pitch-centriccontext (it seems likelythat there are
instances in late 19th-centuryRussian music, perhaps in Debussy, and probably in
Stravinsky[see example below] of pieces most favorablyregarded as having combi-
national diatonic-collection,even triad,centric,but non-pitch-centric syntaxes),I think
it essential to the maximizationof congruitiesamong the relationsbetween the ulti-
matelyavailable properties of a collection (as actually engaged) and the generating
basis inferredfor it and (perhaps) extended beyond it in the "system"that thefunc-
tionallyprimaryelements and relationsbe as far as possible systematically prioras well.
This would insure that the collection- as constructed,or reconstructed- has the rele-
vant hierarchization"built in," as it were, to its "conceptual scheme," as represented
by the sequence and nature of its construction.Thus we will want to constructsec-
ondary individualsas complexes (or subsets)of primaryones (or theirsubsets);and we
will also want to redefine basic relations with respect to their syntacticalfunctions
withinthe system.Thus, for example, afterour constructionof the diatoniccollection
we can construct,by repeated application of the notion of adjacency, the "normal-
form" representation,or scale, after which we can define, for essentialsyntacticrea-
sons, the notion of "scale-degree interval,"whence the "conventional"intervaltermi-
nology arises as a particularsyntacticalinterpretationof constructswhich have more
general (i.e., pitch-class-relational)names as well. It is this scale-degree-syntactical
aspect that accounts for the factthat there are "particular"intervalsof, theoretically,
any general-intervalsize, and that the frequentuse of the scale-degree intervalter-
minologyin descriptionsof nontonal music (or at least non-diatonic-collectional) music
is generallyso awkward.

* 34

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
THE CONSTRUCTION OF MUSICAL SYNTAX (I)

level,each construct[or "theconstruct"]is not regarded as "internally


ordered" or hierarchized, butjust as a complexof discriminablydis-
tinct components); (sub-construct centricities
of dimension-degrees
less than the dimensionsof the principalcollection-partitioning con-
struct,equal to or greater than a dyad; these sub-constructsmay be
regarded as implyinga hierarchizationwithinthe principalconstructs
such that the smallerthe sub-constructdimension,the more primary
its hierarchical position within the principal construct,on the one
hand, and the less uniquely it determines[i.e., the less completelyit
specifies]the total contentof the principal construct,on the other);
pitch(monadic)centricity (the strongestdegree of contentdeterminacy,
comprisingan orderingand hierarchizationof the least-discriminable
elements themselves,in termsof their defined relationsto a single,
referential,such element).
Thus one could perhaps explain much post-tonal"diatonic"music
as collection,or even construct,centric,withoutalso regardingit as
pitch centric. Thus, e.g., a hexachordal segment of the diatonic
collectionis partitionedin Stravinsky'sPetrouchka firstin termsof the
"dyadic construct"(0 2):
TOS: (0 2) (5 7) (9 11)/(5 7) (0 2) (8 10) : T71 (0 = "D")
but later in termsof the "triadicconstruct"(047/037):

(10 2 5) (8 0 3): T3S


which suggestsa local constructcentricitybut a global ("syntactical")
centricityat the collectionallevel only.
Now, with respect to order-determinacy,the necessaryand suffi-
cient condition for its implementationis that sets be comparable in
termsof the variable orderings they determineon a given set of ele-
ments (i.e., thatall setsbe transformableinto all othersby operations
on orderpositionsalone). This requires a dyadic interpretationof the
elementsof everypresented set:
A = {ax y,az,t,. . , am,n}where for all ai,j,i is order positionandj
is a (relative)pitch-classnumber. Then, of course, the same observa-
tions hold for the amount of order-determinacypossible, given only
identity-or-nonidentity-of-order-position information, which be-
comes more determinate under position-relational quantization
("order-classrelations")and stillmore determinateunder durational
interpretation(perhaps like the relation of pitch class to registrally
determinatepitch).
In everysystem(i.e., foreverypitch-timestructure)we can speak of
the degree of order-determinacyrelative to the degree of content-

* 35 -

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
PERSPECTIVES OF NEW MUSIC

determinacy,and note that they are never co-operativeat the same


levels. The corollaryof this is that,by extension,no systemcan be
both content-determinate and order-determinateprimarily. A demon-
onstrationof thisassertionis offeredin the Appendix to thissection.
But first,it should be noted thatthe progressioncollectioncentric/
constructcentric/element centrichas its analog in order-determinate
systems as:

1. determinacyof collection-order(i.e., "12-content"aggregatesas


"partitions"of the pitch-classdomain: chromaticcompletion)
2. determinacyof segmentorder, for segmentsof various dimen-
sions, as:
2 (internallyunordered) hexachords
3 (internallyunordered) tetrachords
4 (internallyunordered) trichords
6 (internallyunordered) dyads
12 monads,
the last constitutingthe maximum degree of order-determinacy.
Note the analogy to our Petrouchka-vs.-tonal music divergencein the
factthat not all 12-tonemusic thatis locallymonadicallydeterminate
is globallyso determinate(i.e., determinatein the same way for any
segment larger than one reference set). Our abilityto discover the
minimum level of order-determinacythat does function globally
should enable us to fixforeach such piece the leveloforder-determinacy
relativeto the(internal)levelof content-determinacy.
Similarly,for tonal
pieces, order of presentedpitchesis an essentialoperativefactor,but,
like content-determinacyin 12-tone pieces, it functions at a less
globally determinatelevel than the deepest level at which content-
determinacyoperates.
2a. APPENDIX
The followingdemonstration illustrates
thenatureof therelation,forany
arrayof between
sets, content-
and and
order-determinacy, indicateswhyin
musicinterpretedas unfolding
within a twelve-pitch-class
octave,the"12-tone
system,"to varyingdegreeof contextualdetermination as to whetherit is
representablein giveninstancesas "1," "2," "3," "4," "6," or "12"-determinate,
is the only possible primaryorder-determinatesyntax(to withininterpreta-
tion of "order" as "registralorder" or "temporalorder," and whereTt and Ir
remain the only relevantsyntacticaloperations):9

9
Althoughthe relationsalso maintainto withinvarioussegmental
determinations
under the "M"-operations
so called and describedin Winham[41], and variously
namedin Howe [16],and Forte[15].

36

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
THE CONSTRUCTION OF MUSICAL SYNTAX (I)

SyntacticalOperations:
If A is a set, and (x, y) is any member of A, then
1. [(((foq(x)), y)= (z, t)) - (z, t) E PqA]
2. - (z,
[(((flq(x)), y) = (z, t)) t) E RqA]
3. [(((x, (foq(y)))= (z, t)) - (z, t) E TqA]
4. [((x, (flq(y)))= (z, t)) (z, t) E IqA]
where foi(x)= (x + i) (mod (s + 1))
and fli(x)= (i - x) (mod (s + 1))
and foi(y)= (y + i) (mod (w + 1))
and fli(y)= (i- y) (mod (w + 1))
and where s is the dimensionof A, and w is the dimensionof the domain over
which,for all members (p, q) of A, q ranges.
Comment: Note thattheserules incorporatethe principlethatwe "compare,"
as "referentialsets"or "macrosets"or "subsets"-i.e., at any given "structural
level"-only sets of equal dimension (as presentedor "canonically"reduced).
Syntactic operations are restrictedto uniformoperationson the entirecontents
of sets of the relevantdimensional degree.
Definitions:
Let A be an ordered pitchsetand (x, y) be an ordernumber pitchnumber couple.
Then,
1. If B is an ordered pitchset,A and B are ordercomparable if,forall y such
that(x, y) is a memberof A, thereis a z such thatfor some q (fnq(x))= z,
where n = 0 or n = 1, and (z, t) is a memberof B, and t is equivalentto y.
2. If B is an ordered pitchset,A and B are orderdifferentiable if A and B are
ordercomparable and thereexistsan (x, y) such that(x, y) is a memberof A
and there is no (z, t) such that (z, t) is a member of B and x = z and t is
equivalent to y.
3. If B is an ordered pitchset, A and B are content differentiableifthereexists
a y such that forsome x such that(x, y) is a memberof A theredoes not
exist a q such thatfor some p (p, q) is a memberof B and y is equivalent
to q.
Comment: All ordered pitch sets of equal token-dimensionor of equal con-
tent-dimensionare content comparable.10
4. Cq is the q-levelsyntacticcontentarrayofA if
A
Cq = {(TqA), (IqA)}
A
5. Cq + 1 is the q + 1-levels.c.a. ofA if
A
CqA + 1= {(Tq +lA), (Iq + 1A)}
6. AA is the maximalsyntactic
content
arrayofA if
AA = {{Cq}, {Cq + 1}, ,Cq + n}} where n = w (see above)
A A

10Content-comparability just requires that every element of one set be derivable


froman element of the other by an operation on thepitchnumber alone (i.e., by pitch-
class transpositionor complementation).

* 37 -

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
PERSPECTIVES OF NEW MUSIC

orderarrayofA if
7. Oq is the q-levelsyntactic
A

Oq= (PqA), (RqA)}


A

8. Oq is the (q + 1)-levelsyntactic
orderarrayofA if
A
=
Oq + 1 {(Pq + IA), (Rq+lA)}
A

9. HA is the maximalsyntacticorderarrayofA if
HA = {{Oq}, {Oq +41}, ... ,{Oq + n}} where n = s (see above, p. 37)
A A A
10. Xp, q is the p, q-levelsyntactic
arrayofA if
A

Xp,
A
q = {{Cq},J
{Cq}}
K
where, for all D and all
r, Cr is the r-levelsyntacticcontentarray of D, and
D
whereJ = (PpA) and K = (RpA)
11. FA is the maximalsyntactic
arrayofA if
rA = {{Xp, q}, {Xp,q + ...XP A q +, n}
n}, 1{X + , q},
A A A
... ,{Xp + m, q + n}} where m = s and n= w (see above, p. 37)
A
12. If FQis a maximalsyntacticarray,then: FQis primary iffor
content-generated
all ordered pitch sets A such thatA is a memberof the syntacticarrayX
and X is a memberof rQ,thereexistsa B such thatB is a memberof S and
S is a member of FQ and A and B are contentdifferentiable.
13. FQ isprimary order-generated if forall ordered pitchsets A and all ordered
pitchsets B, if A is a memberof a syntacticarrayX and X is a memberof
FQ,then,if B is a memberof a syntacticarrayY and Y is a memberof FQ,
then A and B are order comparable and there exists an ordered pitch
set C such thatthere existsa syntacticarrayZ such thatZ is a memberof
FQ and C is a member of Z, and A and C are order differentiable.
14. FQ is uniquelyprimary order-generatedif FQ is primaryorder-generatedand,
forall ordered pitchsets A such thatthereexistsa syntacticarrayX such
thatX is a memberof rQand A is a memberof X, ifthereexistsan (order
number) x and thereexistsa (pitchclass) y such that(x, y) is a memberof
A, then,forall (order numbers)z and all (pitchclasses) t ift is equivalent
to y then (z, t) is not a memberof A, or there is a (are) syntacticalrule(s)
whose application resultsin a set B for which the stipulatedconditions
hold.
From the foregoingit followsthat:
A. FQis primarycontent-generatediffforall ordered pitchsets S such thatS
is a memberof a syntacticarrayX and X is a memberof FQ,and n is the num-
ber of distinctpitch-classelements contained in S, then if m is the total
number of distinctpitch-classelementscontained in the union of all ordered
pitchsets thatare membersof syntacticarrayscontained in FQ,m is equal toor
greaterthann + 1.
Comment: In other words, in order for a syntaxto be pitch-classcontent
determinateat the primary level, the referentialpitch-classcollectionmustbe
a proper subsetof (fail to containat leastone elementcontainedin) the refer-
* 38

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
THE CONSTRUCTION OF MUSICAL SYNTAX (I)

ential pitch-classdomain; and whenever the referentialcollection is such a


proper subset,the syntaxis necessarilycontent-referential at itsprimarylevel.
B. FQ is uniquely primaryorder-generatedifffor all ordered pitch sets S
such thatS is a memberof a syntacticarrayX and X is a memberof rQand n
is the number of distinctpitch-classelementsin S, then if m is the totalnum-
ber of distinctpitch-classelementscontainedin the union of all ordered pitch
sets that are membersof syntacticarrayscontained in FQ,then m is equal ton.
Comment: In other words, a syntaxis primaryorder-referentialif and only
if the referentialcollectioncontainsexactlyone token of each primitivesym-
bol in each place of its ordered couples; i.e., if the element-contentof each
set exhausts the referentialpitch-classdomain, and the token-dimensionis
equal to thatof the referentialpitch-classdomain.
C. There does not exist an ordered pitch set A such thatFA is both primary
content-generatedand primaryorder-generated.
Comment: C followsimmediatelyfromassertionsA and B. With respect to
other than primarylevels of reference, the followingdefinitionsprovide
principlesfor comparison:
Definitions:
15. If for all S such thatS is an ordered pitchset in rA'thereexistsan S' such
thatS' is an ordered pitchset such thatforall y such that(x, y) is a mem-
ber of S' there exists a z such that z is equivalent to y and, for some w,
(w, z) is a memberof S; and, forall (u, v) such that(u, v) is a memberof S'
thereis an (s, t) such that(s, t) is a memberof S and, ifv is equivalentto t,
then if x + i = u thenw + i = s; and, forall (q, r) ifq = n then,if(q, r) is the
lastordered couple contained in S' then thereexistsa p such that(p, k) is
a member of S and p is greater than n, or, if (q, r) is thefirstordered
couple in S', there exists an 1 such that (1, m) E S and 1 < r; thenS' is a
propersegment ofS.
16. If InA is the arrayof all arraysof n-dimensionalproper segments,of cor-
respondingorder position,of distinctordered pitchsets in rA,and every
arraywithinEh is primarycontent-generated,then FA is n-dimensionally
partiallycontent-generated.
17. If Os' is the set whose membersare S' and all othersetsin FAthatcontain
a proper segmentthat is order comparable to S', then,ifOs' containsm
sets, FA is nm-dimensionally partiallyorder-generated withrespectto theS'
array(where S' containsjust n distinctelements).
From the above, it followsthat:
D. If rA is primaryorder-generatedit is n-dimensionallypartiallycontent-
generated forall In where,ifthe dimensionof A is m, n is equal to or less
than m- 1.
E. If FA is primarycontent-generatedit is n-dimensionallypartiallycontent-
generated for all X, where n is equal to or less than m, where m is the
dimension of A.
F. If rA is primarycontent-generated,it is nm-dimensionally partiallyorder-
generated withrespect to all S'-arrays where if the dimension of A is n,
39

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
PERSPECTIVES OF NEW MUSIC

then the dimensionof S' is equal to or less than n -1 and m =(q r) where
q is the number of membersof nA contained as proper segmentsin the
array {S + S} and r is the differencebetween(s- 1) 2 and the dimension
of HA' where s is the dimension of S'.
Comment: All syntacticarraysare either primarycontent-orprimaryorder-
generated, but none is both. An array may be partiallycontent-or partially
order-generatedat given array-levelswhetherit is primaryorder-or primary
content-generated,but never both partiallycontent- and partiallyorder-
generated at the same levels (i.e., with respect to the same internalarrays).
Here followtwo models thatsatisfythe above conditionsfora domain of four
elements,one primarycontent-generatedand one primaryorder-generated.
Domain: {0, 1, 2, 3}
Let A be an ordered pitch set whose membersare (0, 0), (1, 2), (2, 1), repre-
sented below as (0 2 1). Then
AA is:
(TA = (O 2 1)) C0 (IA = (O 2 3))
A
(T1A= (1 3 2)) C1 (I1A= (1 3 0))
A
(T2A= (2 0 3)) C2 (12A= (2 0 1))
A
(T3A = (3 1 0)) C3 (I3A = (3 1 2))
A
and HA is:
(P0A =(O 2 1)) 00 (ROA = (0 1 2))
A
(P1A = (1 0 2)) 01 (R1A= (2 0 1))
A
(P2A = (2 1 0)) 02 (R2A = (1 2 0))
A
and FA is:
(XO, = ((0 2 1), (0 2 3)), ((0 1 2), (0 3 2)))
A
(X0, 1 = ((1 3 2), (1 3 0)), ((1 2 3), (1 0 3)))
A
(X0, 2 = ((2 0 3), (2 0 1)), ((2 3 0), (2 1 0)))
A
(X0,A3 = ((3 1 0), (3 1 2)), ((3 0 1), (3 2 1)))
(Xl,A 0 = ((1 0 2), (3 0 2)), ((2 0 1), (2 0 3)))
(X1,A 1 = ((2 1 3), (0 1 3)), ((3 1 2), (3 1 0)))
(XI,A 2 = ((3 2 0), (1 2 0)), ((0 2 3), (O 2 1)))
(X, 3 = ((0 3 1), (2 3 1)), ((1 3 0), (1 3 2)))
A
(X2,A 0 = ((2 1 0), (0 3 2)), ((1 2 0), (3 2 0)))
(X2,A 1 = ((3 2 1), (1 0 3)), ((2 3 1), (O 3 1)))
(X2, 2 = ((0 3 2), (2 1 0)), ((3 0 2), (1 0 2)))
(X2, 3 = ((1 0 3), (3 2 1)), ((0 1 3), (2 1 3)))
A

* 40 ?

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
THE CONSTRUCTION OF MUSICAL SYNTAX (I)

Let B be an ordered pitch set whose membersare (0, 0), (1, 2), (2, 1), (3, 3),
representedbelow as (0 2 1 3). Then AB is:
(TOB = (0 2 1 3)) C0
B
(IOB = (0 2 3 1))
(T1B= (1 3 2 0)) C1
B
(I1B = (1 3 0 2))
(T2B = (2 0 3 1)) C2
B
(I2B = (2 0 1 3))
(T3B = (3 1 0 2)) C3
B
(I3B = (3 1 2 0))
and nB is:
(POB = (0 2 1 3)) 00B (RB = (0 3 1 2))
(P1B = (3 0 2 1)) 01
B
(R1B = (2 0 3 1))
(P2B = (1 3 0 2)) 02
B
(R2B =(1 2 0 3))
(P3B = (2 1 3 0)) 03 (R3B = (3 1 2 0))
B
and FB is:
(X0,B 0 = ((0 2 1 3), (02 3 1)), ((0 3 1 2), (0 1 3 2)))
(X0, 1 = ((1 3 2 0), (1 3 0 2)), ((1 0 2 3), (1 2 0 3)))
B
(X0,2 = ((2 03 1), (20 1 3)), ((2 1 30),(2 3 1 0)))
B
(X0, 3 = ((3 1 0 2), (3 1 2 0)), ((3 20 1), (3 0 2 1)))
B
(X1, 0= ((3 0 2 1), (1 0 2 3)), ((2 0 3 1), (2 0 1 3)))
B
(XI,B 1= ((0 1 3 2), (2 1 3 0)), ((3 1 0 2), (3 1 2 0)))
(X1,B 2 = ((1 2 0 3), (320 1)), ((0 2 1 3), (0 2 3 1)))
(X1, 3 = ((2 3 1 0), (0 3 1 2)), ((1 3 2 0), (1 3 0 2)))
(X2, 0 = ((1 3 0 2), (3 1 0 2)), ((1 2 0 3), (3 2 0 1)))
B
(X2,1 = ((2 0 1 3), (0 2 1 3)), ((2 3 1 0), (0 3 1 2)))
(X2, 2 = ((3 1 2 0), (1 3 2 0)), ((3 0 2 1), (1 0 2 3)))
B
(X2, 3 = ((0 2 3 1), (2 0 3 1)), ((0 1 3 2), (2 1 3 0)))
(X3, 0 = ((2 1 3 0), (2 3 1 0)), ((3 1 2 0), (1 3 2 0)))
B

(X3, 1 = ((3 2 0 1), (3 0 2 1)), ((0 2 3 1), (2 0 3 1)))


(X3, 2 = ((0 3 1 2), (0 1 3 2)), ((1 3 0 2), (3 1 0 2)))
(X3,B 3 = ((1 0 2 3), (1 2 0 3)), ((2 0 1 3), (0 2 1 3)))
Comment: Note that for all S of dimension n where n is greater than 3, the
syntacticarrayis a proper subset of the arrayof all order permutationsof the
contentelementsof the referentialdomain takenj at a time forall j wherej
is equal to n. Where n is 3 or less, the two setsare equivalent.So our syntactic
operations cannot uniquely determinea syntacticarray for any S of dimen-
sion 3 or less.
From the above, the verificationof assertionsA throughF is unproblematic.

. 41 ?

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions
PERSPECTIVES OF NEW MUSIC

REFERENCES
[15] Forte, Allen: "A Theory of Set Complexes." JournalofMusic Theory,
Winter,1964.
[16] Howe, Jr., Hubert S.: "Some Combinational Properties of Pitch
Structures." PERSPECTIVES OF NEW MUSIC, Spring, 1963.
[41] Winham, Godfrey: "Henry Weinberg: Three Songs." PERSPECTIVES
OF NEW MUSIC, Spring-Summer, 1964.

* 42 ?

This content downloaded from 91.229.248.111 on Wed, 25 Jun 2014 10:37:49 AM


All use subject to JSTOR Terms and Conditions

You might also like