PIERRE BOULEZ Gyno LiGeTt DECISION AND AUTOMATISM IN STRUCTURE Ta If one is to demonstrate the way constructional principles were used in the early stages of serial music, Structure fa’ is a particularly suitable example. Since this com position is very perspicuously worked out, its anatomy is revealed of its own accord, 30 can be analysed as a ‘textbook example’. Alongside the very ramified complexity of the Marteau, it stands in eystal-clear sobriety, At this level of serial technique the Compositional process ean be reduced (0 thee working stages: Decision J~ automatism ~ Decision Il, Decision 1 A, Selection of elements, B. Choice of an arrangement for these elements. C. Choice of the further peratians ta be ea (Carrangements of arrangements’) and mutual relationships of the individuat ‘arrangements to each other. Automatisn Elements and operations, once selected, are, as it were, fed into a machine, to bbe woven into structures automatically, on the basis of the relationships chosen, Decision It ié automatically derived structure is to some extent erude, and one must work on it further, taking decisions in dimensions that are not employed mech- anically. If, for example, the parameter ‘dynamics’ or ‘register’ has not been passed into the machine, then one can work over the erude structure by directing these left-over parameters. This can be done aleatorically, or with definite formal tims, such as to form or avoid particular connections within the given crude structure, ‘ut the above-mentioned division into three results only if, in analysing, one is inclines to simplify. Ie could exsily tempt one to regard the serial mode of working as a dialectic between fieedom and mechanical compulsion, That would be wrong, for in this case decision is not to be confused with freedom, nor automatism withcompulsion. You stund before a row of automata, and are free to choose which one to throw into; but al the same tine you are compelled to choose one of them; you build your ovn prison fs you please, and once safely inside you are again free to do as you please. Not wholly free, then, but also not totally compelled, Thus automatism does not function as the " Mighret hen fr turner he Ui edion ih tere siace a printed eon of the Siracusa ans se counterpole to decision; choice and mechanism are united in the process of choosing ‘one’s mechanisen,? Letus investigate how the process of composition, as outlined, is realised in this piece by Boulez Selection and arrangement of note-qualities* ied down by the given, fixed temperament of the instruments chosen, and also by the ‘waditional twelve-tone method, Boulez employs all twelve available note-qualities. In homage to his teacher he arranges these notes to form the same series as Division | of the note-suecession from Messiaen’s Mode dle valeurs et 'intensités, ‘The marked homogeneity of this series is striking - the frequent occurrence of the interval 11 (Give times), and the lack of 8 and all intervals below 6, The intervals present, apart from 1, are fairly evenly represented: 10 and 7 wice each, 9 and 6 once each, Very sharacteristically, 6 is sed as the final interval of the series (this will late be intensively cxploited in the piece); and the symmetrical position of the two 7s (second and penulti- ‘mate intecvals) is also characteristic, ‘This apparent poverty of the series becomes an advantage, however, since the in- version of the series N Eamte2 consists only of intervals below 7, and thus, apart from 6, has no interval in common with the basis form this makes it ewsier to draw a clear distinction between the in dividual series, which function like threads in a web, Through this arrangement, the tone 6 (Common to the two series) Becomes the axis of the basic series and inversions, while the regions of the remaining intervals function separately in the one case as intervals above 6 (basic series and inverted canerizan), in the other below 6 (aversion and cancrizan of the basic series). This separation is made all the more clearly manifest because the wo triple successions ofthe same interval (I1 and 1 re- spectively) act as characteristic vectors and thus produce a very marked contrast ttween the prevailing moveinent of the two dilferent serial regions. 11 and 1 ulso produce a pronounced chromatic consection within the eres; however, inthe network fvariousl-combined serial threads these connections and prevailing movements ate to a greater oF lesser egies Uestroyed, because between the intervals of one series thele ate tered notes from other series, which distract our attention from the eounections ‘mentioned, snaking us concentsate on other relationships, Thus as the overall structute * fuer ics af etn sue cee fo etbet Ei arite-The Compo’ Fes Ghat, Die Ree fis ube thats ash sad Thess beta Canady BS shy se fenuencies sine the nid ket i eres oon be toaapones te een aeunfolds, there is continually a dialectic - chosen conditions within the series, versus the tendencies (moving in a different ditection) of automatic serial combination. Boulez uses the chosen note-quality series and its inversion, canerizan and canerizan inversion* in all twelve transpositions. In arranging the individual transpositions he telers (0 the following tables.* s 1 TPS SSeTE DON] [T7sWRIINE AES 28 456N 19125700) |7HwWIRe Bt oS 324 34128905672; |Z 1 THe AIO 2SH 45289123 6n107| jo2zN6 S39 KI? Ses omm4n 722 e}29 16 $4108 2738 fies) aioe seen estar) (018 Ges aloes en 10) 71034528126 9) f2 142 ZEIT S 9 e995 6uN7212wW4 13) [oes 27S 4 WIT oe toee) tai @ 109) 3) 2) al) Gs) Oe ee 1 A382, 2210) lox 712824 suo 6 lea 217TH Ss wWIRK oD ln 7azw3 4612958) [82543091 TERM on7t29ate8 s| [sa 827176 IND Baample “These tables result from lining up the original series and inversions according to the note-order of NS, and NI,, thus: [Neo further tables are necessary for cancrizan and canerizan inversion, without more ado if we read the series of numbers from right to lefts 71,1. OF NIC 2, 4,1) ++. 10.” ince these result fy NI, becomes ses we une the flowing bea Bio Original fora Bic faveted aac ‘Selection and arrangement of the (relaive)time-durations “The chosen basic unit (demisemniquaver) is multiplied by from 1 to 12, and arranged in an increasing arithmetical series (again as in the above-mentioned piano piece by Messiaen) yeep oP pL ky oy Et additive structure, behaves heterogeneously as compared with the note-quality sere ‘whose organisation is proportioned.* Ih would, however, be unjust to pass a negative judgment on this dimensional dif- ference, since such disagreements, artistically used, can lead to exciting combinations, Wis mot such a bad idea of Boulez", to link a linear scale (as duration series) to a note quality series that is formed quite differently and is not (or only partly) scalar, One ‘night object more to the unorganic way the durations are permutated. Boulez proceeds 2s follows: since @ duration series DS,: 1, 2, ... 12 corresponds to the note-quality series NSy: 1, 2... 12, a permutation DS,: 2, 8 4, 5... 10, for instance, must cor~ tespond (0 the transposition NS,: 2, 8 4 5... 10, Let us compare the resuiting oration intervals with the note-quality intervals: Bods WPL eh hehe ele hehe) aa Brame Since in a series the decisive thing is the relationship between each element and the ret, the contrast between the wo procedures is obvious; whereas with the (organic) transposition of the note-quallty series, the individual note-qualities are permutated but the interval-relationships always remain the same, the permutations of durations (which die in fact not transpositions at all) occur mechanically, according to tables, and have “Constantly dillerent internal propottions, It is therefore by no means clear why just these permutations have been chosen out of all the possible ones, What is unorganic is this pointless transplantation of a system; note-qualities labelled with numbers, the a»dcmaterialised numbers organised into tables, and the tables finally used like a fetish, as | measure for duration-quantities; thus what were originally mere indications of ‘rrangement are now used as indications of value ‘Selection and Arrangement of the (relative) intensities 'A scalar series is chosen with twelve values from very soft (0 very loud: ees eas eo 10F Na) pppp ppp ppp auasip mp mf quasif fff ft ftir Whereas the predetermined pitches can be exactly realised on the piano, and the durations, ater sufficient practice, rekutively exactly, a consistent performance of these intensity-values is hardly possible. For pitches are ready-made on the piano; durations can be exactly measured (the basic unit can be counted throughout, since, Boulez docs fot Use (00 quick tempi), but dynamics can only be approximately estimated by the performer ~ this is not necessarily a fault, since listeners, too, experience music accord ing to proportions that are subjective rather than calculated. However, itis problematic to uilfferentiate dynamics so subtly as this; the regions of the individual intensity-value, overlap, and one can certainly not be sure, for example, that ap at one point inthe work will not be louder than a quasi p or even'a mp at another point, Thus intensity-values spread out from points to become indistinctly bounded fields, and can only be estimated in eelation to the loudness of their environment (the loudness of notes that are simul tuneous o have recently ceased), ‘Thus one can discern, among the various elemental reions so far discussed, further serial arrangements, according to their degree of pet- formable exaciness 1, Wholly unambiguous pitches, 2) Measured durations, which, however, in the course of the piece are subject to certain variations in tempo, 3. Unmmeasured, only estimated dynamics. ut the serial arrangement of intensities is still more inadequate than that of time durations, Boulez has in fact taken out his chessboard for the purpose, and has pro- ‘seeded diagonally like a bishop: name 7 66 B43) 40 ‘He has thus obtained four arrangements for dynamics: a2 2.7 MoM Ss | s wom 7 7 BR iif, meme CEE quasip quasip fit fife mt mf fit bs 2 2 8 8 Bim 8 8 2 2 5 ‘quasip ppp ppp quisifquasif ffli* flit quasifquasif’ ppp ppp quasip e2o oy t 6 9 7,7 9 6 1 3 2 pep pp opp mp of mf mf mp ppp pp ppp a?oy ot 9 6 2) 2 6 9 F374 mf pp pepe fF mf ppp ppp, mp fpppp ppm (ite search the piece for these arrangements, we shall find deviations at the points frovked *; frinslead of fff, and fff instead of ff. These are subsequent alterations of the dynamic proportions; ia the ection ‘Lent’ ~ bars 32-39 ~the 2nd piano should be tsarkel fff, according wo the higher-order determinations to be discussed later. But this sould be 100 loud in relation to the remaining intensities, and the guast pin Piano 1 Jould be obscured, Therelore Roulee reduced the fit ff. As a resul, he had in the ‘ext section ~ bars 40-47 ~ to prescribe the intensity /finstead of fff, so that the decrease ‘Tone degree in loudness should be retained Thus in view ofthe previous exchange af itor (if, the proportion fff teplaces the orginal proportion fff. Such alterations sre wholy permissible, in view ofthe inistinetness of intensty-values) “Te selection of dynamic proportions according to this diagonal process isineresting ssa gure, butis even less functional than the duration-permutation described; itis not Served from the musical material, but from a numerical abstraction Using the diagonals, one automatically obtains symmetrical successions of values. ‘Te numberof times the various elements appear throughout the piece therefore shows 1 smmetrzal-statistical distribution; two elements (jntensty-values 2 and 7) oocur figntimes, eight elements (13,5, 6,8,9, 1 and 12) four times, and two elements (10) fot at all? The effect of this distribution i that the most differentiated dynamic indiations are used - pppp, ppp, pp, quasi p, mp the simple p, however, never occurs — the only reason being thatthe selected diagonal paths in the tables never touch the figue 4. Choice of timbre: selection and arcangement of modes of attack Boulez behaves very ascetically in this work; the structure proceeds logleally from the schcied elements, and he wants to leave it pure, t0 let ite itsel’, so to speak. The tracure isto emerge os pure shape; for tis purpose two pianos ave particularly well {uted to realise a monochrome network spreading out plastically and transtucently in tine, (The use of vv pianos is determined by the inner construction ofthe music, since the individual serial threads ate woven into two ‘bundles’ that pursue their course fimultaneously.) The sound of the piano is made still more suitable for this kind of ej ea a ea nnn eee Estes 41steucture because after exch atlack there is a characteristic decrease in amplitude; a fact which poiats to the way this music, for all its polyphonic stratification of series, is Formed from discrete elements of pitch and duration: because of the piano's characteristic ‘attack, the acoustic events are concentrated into points that are well defined as wo Frequency and time, The horizontal threads are not indeed unrecognisable, since in th course of the piece they are unmistakably characterised by a constant dynamic ani mode of attack for twelve pitches and durations each; but, all the same, the total ef events consists more’of & structured conglomeration of points. ‘Ehese points are nt ‘geometrical’, that is to say not simple place-indications in the frequency-duration continuum their varying dynamics and moues of attack give them varying weight, lucidity or nuance within the constant tone-colour. Tthis piece Boulez makes use of en modes of attack. It was hard enough to dillerentia:e twelve degrees of intensity, but the distihetions between individual modes of attack ave infinitely subtler, sines piano touch (unlike string instruments’ very differentiated mods of attack which éach produce a quite dilferent formant-spectrum) results from duration. land intensity- proportions that ave always the typical ones, Le., touch is not an in dependent parameter, Ttis obvious that duratious functioning us components of the mode of attack are nota pact of the composition’s serial organisation of duration-values. Durations (predeter mined by the series, of course) are filled out by notes and rests, whose time-proportion is specifi for each mode of attack. Thus with the limiting value legaro, the entire duration is filled out by the note, and the rest thus equals 0; with the opposite limiting value staccarissimo the duration of the note tends toward O while the rest tends toward the maximum, i, toward the serially predetermined total duration; with the remaining modes of attack, the proportion of note to rest lies between those of the two limiting values, without One's being able to fix them defintely (this is to a great extent left to the performer), "Apart from this, some modes of attack have a degree of intensity (such as > or sf) “which still further reinforces the prescribed imensity. This peculiarity of the modes of ick creates difficulties when atlack-scries are coupled with intensity-series. The serial oom works badly When first intensity values and then modes of attack (with their ovn intensity values) ate thrown into the process independently of each other and in such a ‘way that at frst any one can combine with any other. This leads to a number of eon tradictory combinations. For example in Piano 2, bars 48-56, we find ppp poco ‘f, and in bars 73-81 pppp >. Such combinations are most uncertain, and can only be very vaguely interpreted by the performer. But the places with a weak primary degree of ‘tensity are the most problematic ~ the louder intensities are influenced relatively les, since their additional intensity (decided by the mode of attack) is unimportant in com prison with their basic intensity “This ‘counterpoint? between the original intensities and those implicit in the modes of, attack creates fields of inexaciness, which do not follow from the overall conception of the work, 1 was still possible to form a series from intensities, according to a loudness scale But such a scale can not be made distinct for modes of attack, since between the two firmly established limits leyaro and staccatissinw one can not build any succession that will have a clear direction, This is because of the double significance of modes of atiack 2 (duration + intensity), Boulez therefore had to choose an arrangement of the perform- ing indications as he pleased. He arranged the modes of attack as follows: 305 6 7 9 cocoa 2 ‘As we see, the len modes of attack are numbered from 1 to 12, leaving out 4 and 10. ‘This is a result of the arrangement of these elements used in the course of the piece. Correlating with the ordering of dynamics, Boulez uses not the above attack-series as such, but the successions obtained from it by manipulating the diagonals, "Tofind in the serial table the ordinal numbers for individual modes of attack, one has to look for correspondences of certain recurring modes of attack in the piece, and then jook for corresponding number successions in the tabular diagonals. Thus for example in Piano 1, bats 40-46, we find the indication —, in bars 48-56> and, in 37-64 again >; bars 65-72, 7, > and 0, bars 73-81 7, bars 82-89 > and (9; in Piano 2, bara 30147, bars 48-56 in two parts > and in one f, bars 90-97 > and >, bars 98-105 ‘normal, bars 106-115 ‘normal’, > and >>, Starting from the frequently recurring >, itis possible to make out groups of modes of attack, (wo of which are periodic: -. > “=> in Piano | and correspondingly ‘ > Yf >in Piano 2; two other groups, however, Te symmetrica {>A | A> Fin Piano t and = > normal! normal >= in Piano Sin the tables one can find certain periodic or symmetrical number arrangements ‘hich cortespond to these successions of modes of attack, e.g. 1.1.11.15 81. 12.12.16 and 9.1,5.5.1.9 (see Ex, 8). So it is quite certain that the ordinal number 1 ‘marks the mode of attack >, Through a similar procedure one can also identity all the remaining modes of attack, ‘The final result shows four successions: sos uo 3 2 ff 3 mB 8 1 8 4 t= normalnormal = oe a 8 2 2 8 3 5 8 3 5 MW ton od ang normal f+ normal ee 6 6 9 1 5 5 1 9 9 Ay Annet Oe Fe‘These ean be found as follows in the tables y x ena ‘Since in this selection of diagonals two figures are again missing, one sees why Bowles cto works with ten modes of attack, Whereas the intensity series could be made up to 12 by adding p and ff, a similar interpolation into the attack-series is pointless — what doula one insert between “und ‘normal’, or = and -? The enigma of the succession of ‘itacks is farther eomplicated by the fact that two modes of attack are ingerchanged; the Suovession 11 and 3, which occurs twice in the diagonal a, is reversed when it occurs for the second time in the piece": in the section ‘Lent’ (bars 32-39) Piano 2, lowest part (e. F ete), there has to be instead of —, and inthe middle part instead of (because Of the higher order relationships which we shall discuss lates). Since the automatically produced arrangement is no longer symmetrical (a deviation from the succession of intensities), the statistics of the element-distribution are also synmetrical (but always go by even numbers); one element (1) occurs efght times, four ‘ements (3, 6,9, 12)six times, three (3, 8, 11) four times, two (2, 7) twice, and finally the two imaginary ones (4, 10) no times Choice of the arrangement of arrangements “The entice piece is woven from serial threads. Each individual thread consists of one transposition of the note-quality series (wholly strictly, without note-repetitions), in tonjunetion with one of the permutations of the duration-series. All the transpositions Sf the note-quality series are used in basic form, inversion, cancrizan and inverted Ghncrizan, ag age all the permutations of the duration-series (derived by similar methods frown the tables); and each individual form is used only once, This produces a total cf 48 doubly-determined threads. in any particular thread pitches and durations constantly alter as serially deter mined but each thread is homogeneous in its dynamies and modes of attack. These Jauer parameters vary from thread to thread. the individual threads are united to form bundles, $0 that each bundle contains onc, two of thie threads beginning simultaneously (and also ending simultaneously, since 44 ‘every duration series is the same length: 1 + 2+ 3 quavers)” Pianos 1 and 2 each have a bundle running simultaneously, but there are two places where one piano is silent (bars 24-31 and 57-64) so the vertical thread-density of the free escilates between 1 and 6, These double bundles, resulting from the communal Pisa of the two pianos, ate so lined up that there is no time-overlap atthe ‘seams’. “Thus the composition is quite rigidly articulated into sections ~ the place for artistry is inthe balance or contrast of the sections, since within them practically everything hap- fens automatically, As there age fourteen such sections the total length af the plese mounts (0 14 x 78 «= 1092 demisemiquavers. "The higher-order serial arrangement is formed so that the whole composition is divided by a horizontal and a vertical axis, The vertical (ime-)axis divides it into two. In part A the note-qualty series is used in original form and inversion =the duration ies, according to the tabular figures, in cancrizan and inverted cancrizan; vie versa, fart B contains the eancrizan and cancrizan inversion ofthe note-quaity series, and the eration series according to the numerical successions in the originalform- and inversion-tables. "The horizontal axis further articulates these two chronometric sections by dis- trbuting the serial forms (S, I, C, IC) between the two spatially separated instruments, ‘Thus four sectors result, We mark them according to their formal part and instrument sector AI means part A, piano 1, etc. ‘The note-quality series are so laid out that in Sector A 1 the transpositions of the cvigital form occur, and in A.2 those of the inversion, in B | the inverted cancrizan transpositions and in Ti 2 the eancrizan transpositions. The duration series, on the other fund, run in such a way that A I contains the inverted cancrizan permutations, A 2 the ‘Grerizan permutations, B 1 the inversion permutations and B 2 the permutations of the fuse series, Thus the note-quality and duration series are ‘counterpointed’: in A, the tase note series and is inversion are linked to cancrizan forms of the duration series, via vice versa in B, But ths series counterpoint’ is purely external, since, a8 we saw, the ‘urtion permutations have not the same validity as the organic transpositions of the note-quality series. “Thanks to their internal structure, the note-quality series could be distributed with special plasticity. We have seen that NS and NI, of NCand NIC have only one interval ipcommon (the trtone), yet the intervals of NS and NIC, or NI and NC are the same, nly the succession ofthe intervals being reversed, so that the entre serial region falls thre two families. Boulez exploits this in the architecture of the piece: in Piano 1 fist NS, then NIC take their course ~i., only intervals above 6; in Piano 2, NI comes first thea NC, ie., only intervals below 6, This ‘motivie’ separation ensures that although the Rem simultaneous bundles mix as sound, they are to some extent self-sulficient, since they have no intervals in common except the tritone, (All the same this separation, like thet resulting fiom the spatial separation of the two instruments, is not enough 10 lliferentiate the two bundles for the purposes of perception; dynamics and modes of stack have an additional separative effect, as well as thei effectiveness within the ce $12 = 78 demisemi- p 45 tack ta es eh aati ciemeaineechbundles, since they split these up into individual serial threads, Thus there arises a con tinuous dialectical play between separation and mingling of the individual threads and ; in fact this is one of the stimulating features of the work.) tone (common to all forms of the series), which oceurs either at the end (in NS ‘and NI) or at the beginning (in NC and NIC) of the series, serves as an adhesive in the realm of intervals, Boulez exploits this: since in part A NS- and NI-forms are present, ‘each section ends with tritones in the individual threads (this adhesive effect is less powerful if there are more threads simultaneously, because the latter's superimposition interpolates notes from other threads between those of the tritones, thus obscuring them), In part B, where NIC- and NC-forms take their course, each section begins with Iritones. Thus the tritones illuminate, more or less distinctly, first the ends of the sections, then their beginnings, so that they help to organise the overall build-up of the piece. The accumulations of tritones aré produced automatically, yet they result from choice - choise of the tritone as the,series' peripheral interval, choice of the series" simultaneity within the sections, Clearly the automatism of the serial loom can be ‘artistically exploited, if elements and operations are well chosen. Within the four sectors the arrangentent of the serial threads is again determined by seties or tables. In sector A 1, the individual threads are arranged so that the first notes Of the 12 NS series produce the higher-order series NI,; in sector A 2 the total of the. 12.NI series is arranged according to the series NS,, in B 1 the first notes of the NIC series are determined by NIC, and finally in B 2 the first notes of NCare determined by NC,, Thus one can sum up and survey the process: PART A ART Piano 1 | ‘Total of all NS series arranged | Total of all NIC series arrangs according to NI, according to NIC, Piano 2 | Total of all NI series arranged | Total of all NC series arranged according to NS, according to NC ‘These principal arrangements run through the individual sectors without regard for simultaneity or successiveness of threads, This clearly illustrates serial music's general tendency to dissolve the barriers between horizontal and vertical. 'AS we have already seen, the successive sectors (A 1 ~B Land A 2~ B 2) were linked by typical serial forms with common interval relationships. Inthe higher-order arrange. ‘ments, however, these relationships run crosswise; the related nature of NS and NIC links sectors A? and B 1, that of NI and NC links A 1 and B 2. This division is in sccordance with Boulez’ technique of mingling right-angles with diagonals in order to produce an *X-polyphony’. Since there are no connections between the individual duration series such as exist between note-quality series linked by common intervals, Boulez had to unify the uration-arrangements ((o some extent) by choosing the same higher-order series twice in each sector (DIC, in AL and B2, DC, in A2 and B 1); by now it will go without saying, that this occurs in °X' form, 46 element of each durat decided by a higher-order se PART A PART Piano 1 Total of all DIC series arranged | Total of all DI series arranged according to DIC, according 10 DC, Piano 2 | Total of all DC series arranged | Total of all DS series arranged according to DC, avcording to DIC, ‘Yet a flaw has erept into the arrangement ~ though it can only be regarded as a law {tom the point of view of tabular consistency. The IC, sevis runs: 5, 8,4, 6 11, 2,9, 12, 10, 3, 7, 1; but the DIC series built into sector A as a higher-order succession is: 5,8,4, 6, 11, 2, 12,3, 10, 3,7, 1. This is a more radical interference with serial order than the alteration in dynamic succession or the interchange of modes of attack, which have already been mentioned, Here a whole series is missing (DIC,), while another (DIC,) occurs twice. tis impossible to tell whether this state of alfairs isto be regarded simply as a lapse on the composer's part, or as an intentional introduction of *un- predictability’ into the construction ~ a ‘gratuitous action’, as a minor revolt against automatism, Deliberate or unconscious, itis certainly a tiny steuctural fa, but it does sot disturb the pattern (90 miuch ~ sil less since the law is embedded in a tangle of six 4ifferent threads. It is remarkable, though, that in the same section, the first “Lent (bars 32-39), the already-mentioned interchange of modes of attack occurs, apparently torestore the balance; the first flaw" occurs in Piano 1, the second in Piano 2, where one ofthe two loudness-alterations also occurs, Did Boulez perhaps have an off day whet he wrote this section’? In this pile-up of deviations from strict serial order, hardly anything will be percep tile to the listener, so the disturbance of the mechanism can hardly be taken as ubliverate, In strictly regulated composition, ‘gratuitous actions’ are conceivable, even vksirable; by introducing the unpredicted into the predetermined, extraordinary aristic effects could be produced. But an inaudible act of arbitrariness will shock 1io- body except maybe the analyst, And it would be trouble for nothing to build a subllety into the work simply to make a reader rack his brains for an hour or two. ‘An exaggeration of serial thought would here establish a series of degrees of error: 1, faultless; 2. quite small deviation (the dynamic alterations); 3, rather more radical deviation (interchange of modes of attack); 4. very radical deviation (disturbance of the eourse of the durations). But such ‘error-series” could only be functionally evaluated if they were built into the ‘overall architesture in a truly organic way ~ which is not the ease here, ‘Aswe saw, the arrangement of dynamics and modes of attack is not composed with- in the single'threads but only as a higher-order element. Thus in A 1, dynamics are andered according to the succession a (see p. 1), in A 2 according to b, in B | according twe,and in B 2 according to d (but in a and b there are alterations, as discussed). “The arrangement ofthe modes of attack in A 1 is determined by the succession @ (see psd A 2 evording fon B 108 and in B2 10 (wth the exchange alrely nietioned}, ”‘Thus one can summarise the entire plan of the arrangements for the whole structure: PART A PART B Total of NS ace. to NI, Total of DIC ace. to DIC, Dynamics ace. to a Attacks ace, t0 6 ‘Total of NIC ace. to NIC, Total of DI acc. to DC, Dynamics ace. to ¢ Autacks ace. 10 5 Piano 1 Piano 2 ‘Total of NC ace. to NC, Total of DS ace. to DIC, Dynamics ace. tod Attacks acc, 10 Total of NIL acc. 10 NS, Total of DC ace. to DC, Dynamics ace, to b, Attacks ace, 10 ‘This gives a general view of the tabular relationships; the individual threads of the com position are arranged, as 10 the proportioning of their note-qualities and durations, ‘according to the horizontal lines of the tables; the individual sectors are ordered as 10 higher-order note-quality- and duration-proportions horizontally, but, as to dynamics and mode of attack, according to the diagonals ofthe tables; the whole work, however, is laid out, as regards note-quality- and duration-proportions, according to a “diagonal X-concept Iis remarkable that the diagonals for dynamics and modes of attack in the individual sectors again cross, Since the intensity- and attack-arrangement of each individual ‘sector is derived from two different tables (when dynamics are from table S, attacks are from table I, and vice versa), one must imagine the two tables written on top of each other, if one is to see the relationship between the diagonals of intensity and mode of attack ‘The planning of dynamics and modes of attack in the sectors can be sketched as follows: Brample CELT A 48 ‘But the construction, as so far explained, stil tells us nothing about the proportions between the, individual sections, nor about those within the sections, As we saw, the piece is articulated into 14 sections by the succession of double bundles. The individual seotions are made much more recognisable by the alternation of different tempi (as \uration-regulators of a higher order than the duration-series), and by pauses of varying length, But this happens in such a way that the second, thied and fourth sections, and the sixth and seventh, are grouped into two major sections by common tempi and the absence of eaesuras. This reduces the number of sections to 11 (this corresponds ina way to the I-interval series). In the succeeding analysis, to make it easier to follow the overall form, we number the 14 sections Io XI, differentiating the major section 11 into a, Hb and He, and the major section IV into 1Va and IVb. Thus the piece is artic lated as follows: Part A: first section 1 Bars 1-7 second, Ha as third Mb 16-23 fourth, He 2431 fife ML 32.39 sith =) -1Va 40-47 seventh |, IVb 48-56 eighth) V 51-64 Part B: ninth, VI 65-2 tenth =, VI 8 eleventh | VI 82-89 twelfth 4, 1X 90-97 thirteenth, X 98-105 fourteenth, XI 105-115 Thus the distribution of the sections between parts A and B is unsymmetcical ~ a welcome deviation from the otherwise rigid regularity of the construction; parts A and B sue as 8:6 in their number of sections, but as 5:6 if one counts sections and major- sections, Statistically speaking, the horizontal distribution of the vertical thread-density is fairly even, between the limiting density-values 1 and 6: the number of threads present ineach section is: PART A PART 1 Jus} ute [au] ava [ave | v Section wi] vu [van] x] x [xt Number ottheads |2]} 4, afafefa|s fa fa lefelahe“Thus densities 2 and 4 happen thee vimes each, and 1, 3, 5, 6 ties each. The disti- bution is still easier to see from a diagram: Part A Part B 5 Te Po cH te wea a r Teeny t | CaN Tile-b c irva b ¥ VIVUVIIDK XX! In B the average vertical unrcad-dcnsity is visibly greater than in Aj this goes with the we ier number of section, since the total of threads in A and in has to be the same rer ying horizontal distebution of vertical densities in the two parts is particulary we formed; A begins with low density, rises somewhere about the mille to a mari en und falls at the end {0 minimam density; the maximum in the middle js em ved very plastically by minima direty before and after it This sudden concn pees section ITT and the great decrease in density atV have a high degree of surprise io sence part B, just asthe flow of note qualities and durations i the negative of havin past A, a0 the horizontal couse of the vertical dhread-denstis contrasts th rat nthe first part Thus in B the density first decreases fromea higher degree t0 a Tower one, and ends with a steep ascent (0 the maximum, ‘The proportioning of densities is also different in each part. In A we Gnd an ue aymmerial, strongly ossilating density-fow (eg the lap from t to 6 and then back aie to 2) in B there a balance of dnstis: at the begining a decrease from & 10 aa gad reversed and greater proportional change from 2 to 6, The middle valus of auth hese proportions 5:3 and 2:6 would be 4, and in fact we do find density 4 twice penning asa level plateau at the chronometic centre of B.* This products & corte tation between the shaping of density and of the time-flow. ‘ice ihe 14 sections can be rua together ¢o form 11, we can also investigate the Sis: iuibation of the threads inthe light of this organising pattern, In such an investigation vo utatnetion between horizontal and vertical vanishes in sections Ia, b, eand TVs, , tre heteas in al the other seetions the threads run simultaneously, in these evo maior toe ehey run both simalaneously and successively. ‘The distribution of threads in this ease isi PART A PART ® Sections and major sections ve-tia gue succenion tno gu desi dae at nd Yo mana 30 “The distribution ofthe total of threads from sections Ito VI is serial (marked by the ‘racket), The remaining sections IX, X, XI contain threads as indicated by the even numbers in the preceding series, namely 2, 4, 6 ‘alterations of tempo are similarly balanced; of the 14 sections, three ace ‘Lent’, four ave Trés modéré® and seven *Modéré, presque vit". Obviously itis the super-sections Hla, b, e and IVa, b that are marked ‘Modéré, presque vil"; thus their flow of time is brought into proportion with the wo "Tres modéré* and the one ‘Lent’ in part A. In part B the alterations of tempo are more frequent; it changes with each section, This fesults logically; whereas in A the entry of a new tempo is itself endowed with a high Grough degree of surprise, in Tt the degree of suprise must be maintained by more frequent changes of tempo, since one hasalready experienced al the tempi. (One can also tos this from the behaviour of the quickest of the three tempi ~ “Modéré, presque vil ‘tits joins together three sections (Lfa, b, ¢) to form a group, then only wo (Va, B), dnd Jater this tempo occurs only in isolation, in VIt and 1X). ‘The distribution of tempi between the various thread-densitis also has a particular talance, Sections with the same tempo dilfer if possible, in density; ‘Lent’ is combined vith densily 6, 5, 2, *Trés modéré? with 2, 1, 4, 6; only in *Modéré, presque vil" do two Tensities each occur twice (3, 4), and only because this tempo is used relatively more siten ~ the succession of densities for ‘Modéré, presque vil" is thus: 4, 3, 1, 2, 5. 3 4 Inany case, the equal densities in this succession are kept as far apart as possible, and the composer makes sure that they come in a different order, fist 4, 3, chen 3, 4 Boulez has found a very artful way of proportioning the alterations of tempo. To see this better, we use abbreviations for the tempi: L = ‘Lent, M = “Trés modére’, and Ve "Modéré, presque vit" Thus we find these two different symmetrical proportional arrangements: PART A oro ——! —_——> | ——_— A mim Ala a Tl wv Vit Piiv vem] Mik bv Wim i MGV ‘Between the sections or major sections the fermatas are so distributed that out ofthe 10 (pesuras 5 are longer (7) and 5 shorter (P=). With the maximum dierence in tempo {ViL, L:¥), the eaesurais short (>); between the quicker tempi (VM, M:V) itis long {i between medium and slow tempi, however, it varies according tothe order, s0 that ling fermatas ate used when the faster tempo comes first (MM: L) and short fermatas venan the reverse holds good (LM). This is & wholly functional use of fermatas, since when the difference in tempo is greater the separation of the sections isin any case en- Tired, and the new tempo's degree of surprise is higher, so long as it enters without ‘wit delay; on the other hand, smaler tempo alterations do not mark the borders of Heections so clearly, so that longer caesuras have to emphasise the separation (though this only true inthis special ease, where the whole architecture ofthe pice is designed vo thatthe individual sections are to be as easily distinguishable as possible). "Thus we have taken a brief general look at the predetermined construction. The full pate table (p. 52) sketches the buill-up of Structure Ia; here the individual threads rect four-dimensional’ determination, are disteibuted among the individual sections UX stz T ; i ¢ ir : i i BESS ee iar if bect) bes j iE fad ivy ee € 1 a -f ifs ifajee es i ye i © 2 . + | gree [ewe i 1 yee [ee i Z iijee laa : eee lee ; wee i I ] sink i ba i gg Automasison Once arranged, the elements are woven, in the predetermined way, into a network in ‘which the detailed results ean not be foreseen, Since the sections are separated from each, other, the mechanism manifests itself only in their internal structure. ‘The simultaneous threads appear as if photographed over one another; this auto ‘matically regulates the succession of note-qualities. The realm of notes can only be sasped statistically, and one can say no more than that in any section each of the twelve given note-qualities occurs as many times as there are threads, Thus the note- succession is (0 some extent indifferent, and almost every remnant of ‘melody’ is dis- solved (more radically than in Webern, where remnants of ‘expressive melody’, if only fragments, are left hanging from the skeleton of the structures.) ‘The vertical relationships are equally unpredictable, There is no question of ‘chords’ pseudo-chordal phenomena do result when several attacks coincide atthe same point in time, Dut as their components are wholly indifferent these simultaneities have no ‘chordal function ~ they count only as interference-maxima of the threads, with the sole function of greater or lesser vertical density, Thus every beginning of a section is marked by a simultaneity of attacks that has the maximum density specific to that sec~ tion, and this gives a typical physiognomy to the density flow of the entite composition, ‘This uncontrollabilty ofthe horizontal and vertical dimensions, resulting from prede- texmination, has its roots in traditional twelve-tone technique. In the latte, there were already o some extent relations of uncertainty between the dimensions; one could either arrange vertically, n which case one lost control of the horizontal; or one composed with linear series, in which case the vertical simultaneities were very difficult to manage; or else one tried t0 find a common way to direct both lines and simultancities, but had to make concessions both ways, The art of dodecaphonic writing lay precisely in managing the dimensions’ balancing-act in the confined space of the relationship mentioned In our example from the early stages of serial composition the situation gets con- siderably worse; there is no longer even the narrow field of play left by relation of un- certainty ~in fact neither dimension is really available any more. But there is nothing terrifying in this, it only means that the possiblities for composition have moved into sew territory (one previously had not an inkling of the dimensions opened up for com= postion in more recent serial works). So the ‘poor old serilis’ js no more a captive than the tonal composer chained to his eadences; now as ever, the vital thing is how far, and in what way, one can tug at one’s chains, ‘The duration-propottions produced by the superimposition of threads are not to be influenced any more than are the successions of note-qualities. Variations in horizontal density (frequency of attack) ~ horizontal alteration of the vertical attack-density: both are now taken out of the composer's hands, Let us take a few examples, First the superimposition of two threads in section 1: wp Roop ype pop ay od? pee py oy oe es ss LULU ample 1 3In seotion 1Va, four threads are simultaneous: yy ae papa oe By pho RE vend oP am ee a ' A oy DARA J Jb eh , Fora still more complicated case, take section IIL, For simplicity’s sake only the sum total of the six superimposed threads is quoted: Apap mean a ae pena “Thus the adding together of regular processes gives a wholly irregular result, of very elegant flexibility. The fewer the threads working together, the more richly varied is the flow, since no duration value occurs much more often than any other (in this case, how- ever, there is still not too much damage to the original layout ~ all duration values once each, maximum degree of variation, as seen in the single-thread sections He and V). But when the number of threads increases, the number of smallest phases (demisemi. ‘quavers) increases very rapidly, and the frequency of semiquaver values also increases, it less, stil fairly rapidly, Both multiply at the expense of longer values, which slowly disappear, This can be seen from the comparison on opposite page. It is not by chance that the highest thread density in the piece is 6. Ifthe threads were to pile up any further, the demisemiquaver would devour the longer values so complete} as to produce @ boring, almost periodic pulsation, which would rather harm the festhetic value of the piece. This is why the already-mentioned ‘flaw* in the flow of ‘durations (in fact it occurs in the dense section II!) has not the slightest disturbing effec: ton the overall picture, since demisemiquaver values are so frequent. ‘The horizontal variation in vertical attack-density is still more irregular." As we saw, this density, determined by the nature of the construction, reaches its maximum at the beginning of the sections. Elsewhere, ‘the density obviously rises as the number of threads increases, but where and how many attacks coincide is left to chance: an inter section of separately determined processes, which from a statistical point-of view dos not happen all that often. ‘The extent to which the individual threads in the web are distingwahable could also be regarded as a mechanical end-product, Here three factors are determinant: dynamics, 2 See Baty Hand 15 % mean 99 hes, IC Hor 3K ae sano ach 54 ‘Number of values per section vane (ovo seats) | courted). | (ix ead 3 3 2 3 e > | * 4 oa 2 2 2 glk 4 1 1 é. 3 1 1 L 1 2 ° ode of alack and the numberof treads present, Manimum clarity i obviously ‘aie by the singletnead structure in section Ile and V. But the fist section, to, ton bovery cal ‘heard thou two teas wi vty beat itetends in dynam ‘Uff and quasi p). In section X the two threads are rather less plastic; although the dis. ction in mode of attack greater, the dynamic distinction are much ss iid tan in Of the theeavead structures, section VIL can be clearly heard through; two leads which in fat hae the same mode of attack (>) are composed witha vty great sifltence in intensity ppp), an the third (np) is very past because fis mode of aac (legate) Amini of separate existence or te individual heads i found in eset nal esl hx sta one sg i pled mor a oe 2), lo fily homogeneous dynamics (vo threads each pop and py, one each ani np. Among the two-thead stractrs, oo, there is one tat fay welded to paer: ction IVa, played ff and quai The sttustures ofthe other sections dr paste or inistnc to varying degrees, as stages between the etteres mentioned. Decisions within the products of automatism ‘These occur, as discussed, through choices in some parameter not yet exploited; in this ease, register-disteibution of the notes, and, in conjunction with it, the direction of their movement, Note-qualities only become pitches when their register has been determined; this changes an abstract ‘quality’ into a note that ean sound, We have seen that intervals, both built up vertically and lined up horizontally, are ‘ail indifferent in this work, But there is one exception: the serial structure is hyper- sensitive 10 octaves. One is familiar with this from traditional welve-tone procedure." There is something vague about the octave as an interval; it is not a genuine note Ne ed Aan sur Zvontonpcion Visine 1932 (Univeral Edo), Cape 6: Die Ode 35repetition, yet it mingles too much with the original frequency to be peresived as wenuinely different. In choosing registers Boulez to a great extent follows this principle of avoiding ‘octaves, This leads to serial degrees of freedom of choice; in the monodie sections, the repister-distribution is free; the more simultaneous threads are present, the more likely are octaves to occur, and the choice of registers becomes increasingly fixed, to ensure that they are avoided as far as possible “This is a particularly good example of the way decisions hang togethey; the original choice of vertical thread-density for each section affects the free choice of registers, because of the principle of avoiding octaves. Thus one choice inevitably imposes limits fon all the decisions that follow from it, “The avoidance of octaves means the piling-up of note-repetitions ~ these occur because the chosen threads are woven fn fixed registers, These note-repetitions again produce very plastic connections within the structures; so that further patterns are built into the irregularity of the network; this dialectic between the apparently regular and the seemingly accidental is in fact one of the most attractive characteristics of the piece, Let us examine @ few sections from the standpoint of register-distribution, For clarity’s sake we run all the notes of each thread together, ignoring the time-dimension; section I then shows the following distribution, which is almost ever Ns. Os We sce that nine note-qualities ~ i,, the great majority ~ are in registers common to both series; only theee (c, e, g) differ in position, This means a greater degree of fixing than would have been necessary to avoid octaves." This leads 10 frequent note- epetitions, which attract our attention when we hear the piece. The resulting con- nections have the effect of knots in the structural network. These knots are all the tighter as fewer attacks are inserted between a note and its repetition. (These repetitions in the same register are thus chosen with relative freedom; but their degree of separation is awlomatically predetersnined by the nature of the serial weave), & Eramgle 13 2 Taina nots io every asin; frenatpl Im eton IVs eich aio onopat ony Bre pie ae conn 56 Ween lista series of degrees of connection, which emerges clearly from the structure, For this purpose we sketch the course of Section I, without indicating the rhythm, ‘The degrees.of connection, arranged in a scale, are as follow: Degree of connection O* :¢'""~C ; B-e'"";G—g'""" Smallest actual degree of connection (1): B-B (14 other pitches between) Degree of connection 2: low B Flat ~ B Flat (12 other pitches between) 35 £409 other pitches between) 4: d= d" (5 other pitches between) hom) SHOW A FlateA fla (4 other pitches) homo 64 za’ —a! (one foreign attack between) 74020" sharp ~ {°" shagp (one foreign attack between) ic" sharp —c'* sharp (direct succession) og rel!” flat e'""’ flat (simultaneous) + This marking means that these pairs are far apart not only in register but in time, to that no trace of a connection is present. One can only determine the degree of connection between these pairs if one takes into account how near in time the note is to its repetition, ‘Thus the interval of entry of, the fsharps is quaver, that of the as isa erotchet plus semiquaver, so the first pair has 1 somewhat higher degree of connection +++ In fact the maximum degree of connection is not a connection at all (maximum ‘quantitative alteration is transformed into something qualitatively different), because an ‘ement is absent ~ recollection of pitch that has already sounded x0 there results not ‘a‘conneetion’ but complete welding together. In any case note 1 of NS, swallows the first note of NI, since it is struck as against the quasi p of the other, and also lasts longer. If we arrange the individual conneetions chronologically, taking the points where the connection is perceived, ie., the second note of each pair, the succession of degrees of connection is as follows: 9,64, 3%, 82.31 (and 5 struck simultaneously, but 7 more prominent than 5). is no accident that the lower-degree connections are left tll near the end of the section. As time passes, it becomes more likely that these looser connections will occur, since they take more time than the higher-degree ones (thus most of the latter are shifted forward toward the beginning) ‘This series of *knot-strengths’ shows that at the beginning the threads are woven. together particularly firmly; the web then loosens fora time, only to tighten again in the siuth bar of this seven-bar section (our attention is especially caught by the repeated sharps). Then the web quite suddenly Joses all its tension, since the note-repetitions at theend of the section lie so far away from the first appearance of these pitches that they are hardly to be perceived as repetitions. Since there is a great dynamic difference between the two threads, the note~ repetitions act at one point as a note with an echo (/f'~ quasi), at another as a ‘pre- echo’ of a note that actually enters later (quasi p ~ ffff). This enriches the structure through stereometric eflects of nearness and distance, ‘The fixed register-relationships are more complex in section Ifa, where four threads are interlaced, 2Here repetitions occur not metely once but two and three times. These multiple note- repetitions ure particularly stimulating because this section is highly differentiated in its dynamies and performing indications. Notes repeated three times (i. sounding four times) are; os) g Exnple 17 “The note b fat frst appeats in the section's fist bar (bar 8), ppp —- in Nl, then (already almost forgotten because of the interpolated repetition of the note g) it returns in NS,, mf legato (bar 10), and immediately afer (a semiquaver later) as an immediate echo in NIy (pep, normal). A dotted minim plus semiquaver after this, itis suddenly struck as a sharp mf (NS,, bar 13). The note g behaves similarly. However, a more interesting and conspicuous repeti tion is that ofe fla, since by the automatic workings of ‘chance the four e flats come very close together. In the first bar of the section (bar 8) we frst hear © Aat ppp (in Nip), and in the same bar, u dotted semiquaver later, i reappears louder in NS, (n/, Jegaio), as if the intially thin thtead had suddenly thickened; in the next bae (dotted «quaver later) the process is reinforced still further: € flats occur simultaneously in both INS, und NI,, and the one in NS,, struck sf, rounds off this succession of repetitions very pointedly. Thus the note e flat becomes especially conspicuous, emphasised by the recollection of the e fat three octaves higher at the beginning of the first section, But ‘one may not regard this note as a "tonic" or central note - nothing ofthe kind can exist in this kind of musie since to compose ‘serialy* means the abolition of any hierarchy of the musical elements); the pile-up of this note, being a purely accidental result, has no harmonie function The three threefold repetitions are placed close together in the middle register, whereas the four twofold ones le wide apart in the Frequeney-spa eapte 8 38 OF these es, the pitch f sharp is particularly emphasised, being almost of equal importance with the repeated e flats ~ moreover its appearance corresponds practically symmetrically. to the latter, coming almost at the end of the section (bar 15). Thus section Ha is held fiem as if with a nail at either end. ‘These {sharp repetitions occur as follows: frst ppp in NI, and immediately after (a éemisemiquaver later) mf in NS, and NS, simultaneously, reinforced by the com- bination of 4 and legato, ‘The other double repetitions are less important than this, but still noticeable, The low F toward the middle of the section (bars 11-12-13), and the b and ¢ toward the end (burs 14-15) produce, together with the f sharps, a whole web of note-repetitions, OF the single repetitions, ¢ sharp is very pregnant (in NI, and NS,, bar 11), because the wo identical pitches follow each other very closely (phase-diflerence of a quaver). The second ¢ sharp (NS,), because of its intensity, and especially because of its mode of attack (sf2), emerges specially audibly from the three-note simultaneity, since the other two notes (E from NI, and f sharp from NI,) ate to be played ppp. ‘The remaining single repetitions have less effect than those mentioned, ‘These and similar connections occur in all sections, and each section has it network of nots, denser or less so. We saw that, in order to avoid gctaves, the fixing of registers intensifies as the number of simultaneous threads rises. This increases the number of connections, and the sections with greater thread-density are also denser from the point ofview of note-repetitions. Thus the density of the connections is at its maximum in the siepart section 11, ia which all egisters are xed: common 0 85 NS NSy Nig Ny My & & Banple All the connections produced are six-fold, so that they produce a wholly static result, and they have a particular charm, nuanced as they are by intensities and modes of attack, Since there isa relation of uncertainty between the choice of connections and that of the parts’ interval directions, the complete fixing of registers means that the composer «an no longer influence the course of the individual threads, which is wholly automatic, ‘The conditions in section XI are similar to those in LE, with one exception; there are only 11 sixfold connections ~ one note-quality (B) changes register, Thus one ean draw upa scale of degrees of connection, from the maximum connection in section II to con- section nil in the two monodie sections (IIc and V); this scale isa series ofa higher order dian the connection-series within the sections, But within the mult-thtead sections relationships of other kinds can be heard; par- ticular formations of a motivic kind, which (as we have already indicated) are no longee really motives but something like the shadow of the remains of motives. Sucl forma: ‘ions occur where particular intervals ate repeated, The occasion for this is usually one ”‘of a number of relationships between the individual note-quality series, These relation ships are partly.predetermined by automatism, but ean be emphasised with varyin, clarity by the (limited) choice of register One opportunity is provided by the tritones, which have been mentioned as the “end. intervals" of the series, We said that forms NS and NI, NC and NIC have no othe, intervals in common, because of the way the series is constructed. There are all the more intervals common to the various transpositions of the same basic form. Obviously the ‘tucture could not exploit all the possibilities offered by the combination of series, Ang Boulez clearly did not steain himselt 1o pile up these relationships, since if there are toc many ghosts of motives they materialise and become real motives; this coagulatior could disturb the glassy transparency of the network, To permit such relationships, the registers of the intervals must remain the same, ane the return of the interval may not be delayed beyond the time within which recognitioe is possible In Section Ia, for exainple, apart from the tritones mentioned, four other interval ‘epetitions would be possible: a ~a Nat and d ~¢ sharp in NS, and NS,; ab flat anc =f in. NJ, and NI,. But only the latter is exploited, with marked effectiveness in the bass; E and low F, both times ppp, first with a short attack (_), then normally (bar, 10-11 and 11-12), Plasticity is increased by the return of low F mf in bar 13 (sce the fow-F connection mentioned). A still more striking passage, which stands out of the structure almost as a normal motive-imitation, ean be found in section IIb: the co- inciding series NS,, und NS,, have a three-note succession in commun (€~e lat ~ This complex is reinforced by the identical registers of each pair of repeated note, (bars 17-1 Emule 20° ‘tn bac Saye top parte semiguaner set i er hee at, ‘his comes close to being a traditional ‘augmented imitation in stretto’, yet this passage thas a quite different effect froma similar one in earlier music, In the latter, high degree: of surprise are usually associated with the appearance of elements not previotsly used ‘but in serial music of the kind investigated here, one result of the total exploitation of ‘note-qualities is that the sudden literal repetition of a several-note figure in fact gives the impression of something wholly unexpected, and thus has a specially high degree of surprise 11 is interesting 10 observe how in very dense webs, e.g., in section I, motivie relationships are obscured even though their notes shate the same registers, For the ‘multiple note-repetitions attract our attention, preventing us from perceiving the rest ° 0 the interval-relationships, For example, the interval f sharp - F (bars 33-34, NS, and NS) is also imitated in stretto, and the two F-s even follow each other with a phase- uillrence of a mere demisemiquaver; but this process is strangled by the labyrinthine network, and is also obscured by fff-attacks. It is of no avail that the same interval £ sharp ~ F has already been heard at the beginning of the section (bar 32, indeed not in te same thread, but as the first note of NI, and the second of NI) Such motivie ‘elationships, which do not result from a serial relationship but draw their notes from several simultaneous series, often occur when the constellation of the threads is favour- able. A particularly pregnant example of this kind is found immediately after the teyinning of section Ifa (bars 8-9). We have mentioned that in this passage the pitch «flat stands out. But if we also examine the passage from the point of view of interval- imitation (one does not even have to look ~ one hears it the very first time), we find a \ery surprising example ofthe way motives are built [con all four series (g sharp ~e flat initated) = So when we hear this composition a complex network unfolds ~ of coarser or finer weave, variable aural perspicuity; consisting of a significantly ordered flock of sounding ‘poins': these are organised to form threads of varied thickness, which now stand out plastically, now become less distinct, The threads, for their part, are woven together with greater or lesser density. For the listener, knots, relationships, connections of many Linds emerge; the result isan organism as ramified as itis elastic, We have seen that this organism is the result as much of decisions on the composer's part as of automatic nnechanisms; and that these decisions and automatisms are not opposed principles but |woaspects ofthe samme principle. Interacting decisions ead unavoidably to automatism, deiemmination creates the unpredictable; and, vice versa, neither the automatie nor the secidental can be created without decision and determinin, Seen at close quarters, itis the factor of determinism, regularity, that stands out; but seen froma distance, the structure, being the result of many separate regularities, is seen tobe something highly variable and chancy, comparable to the way the network of neon. lights ashes on and off in a main street; the individual lamps are indeed exactly cou- trolled by @ mechanism, but as the separate lights flash on and off, they combine to forma statistical complex, ij aScen froma greater distance, these light-complexes merge (0 form @ higher unit, ohigh is also significant in its own way; So, 100, does the structure of this musie when curd often, Our perception, which at first only noted the accidental details, then ‘iucieates gradually to deeper levels, tll it discovers the overall coherence and pro- rortions, Its just these latter that give the piece its artistie value, which can hardly be rusped by listening in @ traditional way. The *beauty" of piece like this lies in quite ‘ew qualities, Webern's interval-objects, as we said, still contained a trace of the (dis- rectly) ‘expressive’, and although the satisfaction derived from his music is the result Of quite different qualities, the traces of ‘expression’ present at times do provide Crutches for the struggling listener, All this has vanished in our example from Boulez! ‘Structures; they expose to view something that in Webern already formed the nucleus; beauty in the erection of pure structures: = ‘Since in music this ean only be achieved through time, composition at the serial level as become work with time, Thus composition ceases to be essentially ‘art-work’; wo Compose now takes on an additional character of research into the newly-discovered Telationships of matetial, This attitude may strike people as negative, ‘inartistic? ~ but tdiete is no other way for the composer of today, if he wants to get any further. For troulea, basic experimentation of this kind produced Structure Ia, In 1b, dry severity is rlaxed a little (relatively free splisting-up of the determined duration-elerients), in eit again more Figid. ‘This ascetic attitude, akin to compulsion neurosis, is self-limitation from ehoics ~ as if dhe composer Were taking himself for a walk on the end of a lead; yet Bolulez had to break away from it, ix order to thcow himself into something completely opposed (but internally related), And so he ereated the sensual feline world of the *Marteaut. a