ILLINOIS UNIVERSITY OF 1LUN01S AT URBANA-CHAMPA1GN PRODUCTION NOTE University of Illinois at Urbana-Champaign library Brittle Books Project, 2014.COPYRIGHT NOTIFICATION In Public Domain. Published prior to 1923. This digital copy was made from the printed version held by the University of Illinois at Urbana-Champaign. It was made in compliance with copyright law. Prepared for the Brittle Books Project, Main Library, University of Illinois at Urbana-Champaign by Northern Micrographics Brookhaven Bindery La Crosse, Wisconsin 2014UBRARV OF THE ^MMlWTirWMm THE MOTION OF THE VOICE IN CONNECTION WITH ACCENT AND ACCENTUAL ARSIS AND THESIS. By C. W. L. Johnson. BALTIMORE, 1902.Reprinted from Studies in Honoe of Basil Lanneau Gildersleeve. THE MOTION OF THE VOICE IN CONNECTION WITH ACCENT AND ACCENTUAL ARSIS AND THESIS. The fact that there is in all articulate speech an element of pitch needs no proof. It can be observed in every modern language. Its existence could be assumed for ancient Greek and Latin, even if there were no recognition of it in the writings of musicians and grammarians. As a matter of fact the presence of pitch in the tones of the human voice was considered of sufficient importance by many Greek theorists to warrant a formal analysis of the manner in which variation up and down took place. Our chief authority for this analysis is Aristoxenus of Tarentum. In his harmonica elementa, I, §§25 if., p. 8 Meib., p. 10 Westph., vocal motion is divided into two classes, the continuous (κίνησις συνεχής) and the intervallar (κίνησις διαστηματικη). In the former the variation in pitch is such that the passage from one degree of pitch to another is through all intermediate degrees, and the pitch is nowhere stationary for a perceptible interval of time. In the latter the passage from one degree to another is by a leap, so that there is no fluctuation during the production of a note, but the pitch remains steady now at this, and now at that, degree. These two forms of motion characterize the speaking and the singing voice respectively, and the analysis of the pitch-changes seems to have been made chiefly for the purpose of differentiating these two kinds of utterance. Aristoxenus expressly identifies con- tinuous motion with the variation of pitch which takes place in speaking, and intervallar motion with that which takes place in singing (harm, elem., I, §28, p. 9 M, p. 11 W).1 Thus a comparison was instituted between the two most important forms of human utterance, speech and song, and the 1 Later writers make the same or a similar classification : Vitruvius, de archit., V, 4, 2 ; Aristides Quintilianus, de musI, iv., p. 7 M,p. 4, 26 Jahn ; Cleonides (Pseudo-Euclid), introd., 2, p. 180 KvJ; Gaudentius, introd1, p. 328 KvJ; Claudius Ptolemy, har??ion., I, iv., p. 8 Wallis; Martianus Capella, IX, 937 (318 G).58 C. W. L. JOHNSON. melodic or tonic element in the one was considered in connection with that in the other. Variation in pitch is common to both ; it is the manner of the variation which is different. A succession of fixed pitches, that is, of musical notes, subject to certain rules in regard to the width of the intervening intervals, constitutes the melody of music proper. A succession of fluctuating pitches, while it may not conform to so definite rules, nevertheless pre- sents a no less interesting phenomenon. Such a succession Aristoxenus calls λογωδίς η μέλος, "a conversational melody" (harm, elem., I, §42, p. 18 M, p. 17 W), Cicero, cantus obscurior (or. 17), Dionysius of Halicarnassus, διαλέκτου μέλος (de comp. verb XI.), ro της φωνής μέλος, λέγω δ' ου της ωδικής άλλα τής ψιλής and τα μέλη τών φθόγγων (ibid., xi., fin.). The word προσωδία and its Latin equivalent accenhis imply the same conception.1 The nature of such prose tunes will depend upon many con- siderations. Every language has its own characteristic forms of melody, every individual speaker his own variations on the national air, if I may call it that. Statements have one form of melody, questions another. The various emotions, anger, com- passion, hatred, contempt, and so forth, find expression in the tune which runs beneath the words. In many languages it would seem a hopeless task to formulate the laws which govern pitch- changes. Laws there must be, if they could but be unravelled. But in the case of ancient Greek, at least, the differences in pitch were so marked in point of size and so uniform in occurrence, that a formal classification of the variations could be made. Long before a system of written accent-signs was devised, the stable character of the melodic outline of Greek words as pronounced in ordinary speech was recognized. The pitch element in words was, it would seem, almost as much a fixed characteristic as is the stress or intensity element in English. The pronunciation of the individual speaker might present variations (in degree), but each 1Diomedes, p. 431, 1 Keil : accentus est dictus ab accinendo, quod sit quasi quidam cuiusque syllabae cantus. apud Graecos quoque ideo προσωδία dicitur, quia προσόόεται ταϊς σνλλαβαϊς. Servius, de fin., p. 451, 10 Κ : accentus autem est quasi adcantus dictus, quod ad cantilenam vocis 110s facit agnoscere syllabas. Martianus Capella, III, p. 65, 19 Eyss.: et est accentus ut quidam putaverunt anima vocis et seminarium musices, quod omnis modulatio ex fastigiis vocum gravitateque componitur ideoque accentus quasi adcantus dictus est.ACCENT AND ACCENTUAL AXSIS AND THESIS. 59 word within the dialect had at any given epoch a normal scheme of high and low pitches, to which the pronunciation of all those who spoke the dialect tended to conform. The existence of a tonic or melodic accent in the Greek language throughout the classical period and down to Roman times has been abundantly proved. It is not necessary in this connection to review the evidence on which the universal con- viction of scholars on this point is based. Besides the testimony of Aristoxenus we have that of Plato and Aristotle.1 We learn from Varro that Theophrastus and Eratosthenes were interested in the melodic accent of their language.2 After the invention of the written accent marks by Aristophanes of Byzantium the sub- ject of correct accentuation became naturally more and more important from the point of view of grammar, and the theory of the accents was handled by many writers. Dionysius Thrax taught that there were three accents, the three which became universal, acute, grave, and circumflex. Clear indications of the nature of the Greek accent are contained in the de compositione verborum of Dionysius of Halicarnassus. He states at the begin- ning of chap, xi that prose diction to be artistic must attend to these four things, μέλος, ρυθμός, μεταβολή, and τό πρέπον, and in the middle of the chapter that variation in pitch takes place within the compass of the interval of the fifth. The whole of a word is not spoken with the same pitch, but one part with οξεία τάσι?, another with βαρεία, and another with both (one after the other, of course). In chap, xix he speaks of the accents as τάσεις φωνής ai καλούμεναι προσωδίαι. The melodic element in the language was evidently far more important to literary and grammatical studies than any intensity or stress element of the sort found in most modern languages. Differences in intensity cannot but have existed, but in the absence of any formal consideration of them by ancient writers, it is reasonable to hold that intensity-variations were always affections of the whole sentence and not of individual words as such. The parts played in modern English by pitch and intensity would thus be reversed in ancient Greek. While in English stress concerns chiefly the pronunciation of individual words (although there exists an important sentence-stressing 1 Plato, Crat.y 399 A-C ; Arist., rhtt., Ill, 1, 4. 2 Varro ap. Serg., de acc., p. 189 Wilm. (Keil, Gr. LatIV, p. 530).6o C. W. L. JOHNSON. superimposed upon the succession of word stresses), and pitch- changes affect the sentence as a whole either as an oratorical element or as a capricious manifestation of personal taste; in Greek, on the contrary, pitch concerned chiefly the pronunciation of separate words, and any oratorical effect produced by pitch- changes was effected through a superimposed melody demanded by the emotional character of the sentence as a whole, in the same way that oratorical emphasis is a concern of the whole sentence. When we turn to the accentuation of the Latin language, we are not surprised, in view of the work done by Greek grammarians for their own language, to find that there exists a great mass of writing professing to deal with the.corresponding phenomenon in Latin. But, whereas it is now all but universally conceded that the Greek προσωδίαι were, what their name implies, semi-musical affections of words, there is not among Latin scholars the same unanimity in regard to the true nature of the Latin accentus. One party holds the view that in Latin of the classical period at any rate the verbal accent was essentially the same phenomenon as was observed by the Greek grammarians in Greek speech. The other party, now in a majority, would make the accentus an intensity or stress accent of the same general character as the accent in modern English and German. It is not my present purpose to offer any argument for or against either of these views, but it must be admitted by any one who will read the passages bearing on the subject in the writings of such authors as Cicero, Varro, Vitruvius, and Quintilian, to mention no writer of later date, that, rightly or wrongly, these authors thought that there existed in their language a verbal melodic accent, strictly comparable to the Greek accent. They may have been mistaken. If so, we may reject their evidence in reaching a decision as to the true nature of the Latin accentus. But even so, it is important to analyze the delusion under which they suffered, if only for the purpose of appreciating just how far it extended, and just how far it invalidates their evidence on other questions closely connected with that of the accent, as for example the metrical question. We must become alienists for the moment and for the purpose. Evi- dently the whole truth can never be reached if we confine ourselves to etymological and historical considerations, much less to those which are evolved from our inner consciousness. It is imperative that we regain the ancient conception of the matter, if we intendACCENT AND ACCENTUAL ARSIS AND THESIS. 61 to make even the slightest use of the doctrinal matter which the ancients have left us. Now assuming for Latin a melodic accent, real or imaginary according to our preferences, let us consider the phenomenon presented to the ear by ancient accent in general, and its treat- ment by ancient theorists. For this purpose it will be convenient to imagine the changes in pitch values to be represented by a line traced by a moving point, in such manner that its motion from left to right denotes the passage of time, and its variation upward and downward the variation of acuteness and graveness. In the first place there are only two possible directions, up and down, in which variation can take place. So long as pitch alone is under consideration, there is only one dimension for the move- ment. But the number of degrees which may be recognized in any system of denoting pitch is limited only by the ability which the inventor of the system may fancy he possesses to discriminate with certainty the finer grades ofpitch. Theoretically there cannot be too exact a notation to denote the subtle gradations and varia- tions of pitch easily detected by the trained ear. Continuous motion demands a more complete notation, if it is to be scien- tifically recorded, than does the intervallar motion of music proper. In practice however the more complicated the system of notation, the more easily will it break down. If the more striking variations from the mean tone of the individual voice are indicated, a sufficiently accurate record for practical purposes would seem to have been devised. Another consideration bears upon the kind of motion involved in ordinary speech. If the definition in Aristoxenus of the con- tinuous and conversational motion conforms to the facts as observed (and we have no reason to suppose that it does not), there are, strictly speaking, no stationary pitches at all in this form of motion. Says Aristoxenus, harm, elem., I, §26, p. 8 M., p. 10 W.: " In the continuous movement the voice appears to the senses to traverse a certain space in such a way that it rests nowhere, not even, so far as our conception of the sensation goes, at the bounds, but is borne along continuously until the sound ceases."1 And a little further on he says, ibid., §28, p. 9 M., p. 1 Aristoxenus, harm, elem., I, §26, p. 8 Μ : κατά μεν ovv την συνεχή, τόπον τινά όιεξιέναι φαίνεται ή φωνή τη α'ισθήσει, όντως ως αν μηδαμοΰ ιστάμενη < y >, αηδ' £7γ' αυτών των περάτων, κατά γε τήν τής αισθήσεως φαντασίαν, αλλά φερομένη συνεχώς μέχρι σιωπής.62 C. W. L. JOHNSON; ii W.: "Now the continuous movement is, we assert, the move- ment of conversational speech, for when we converse, the voice moves through a space in such a manner as to seem to rest nowhere. In the other movement, which we call intervallar, the contrary process takes place. For the voice seems to rest at various pitches, and all say of a man who seems to do this, that he no longer speaks, but sings. Therefore in conversing we avoid having the voice rest unless we are forced at times by reason of emotion to resort to this style of movement; but in singing we do the reverse, for we avoid the continuous and strive to make the voice rest as much as possible. For the more we make each of the sounds one and stationary and the same, so much the more accurate does the singing seem to the senses. It is fairly plain from the above that of the two movements of the voice in respect to space, the continuous belongs to con- versational speech, the intervallar to song."1 Now evidently a notation would be severely taxed if it attempted to indicate all the glides characteristic of our conversational speech. Not only are the bounds of such downward and upward movements difficult to determine from the nature of the case, supposing it to be true that all speakers employed exactly the same glides for the same words, but also the rapidity of the ascent or descent would defy accurate analysis. A sentence in Greek, then, presented—what is seen in every language of which we can to-day study the actual sounds—a complicated succession of glides in pitch, some of them short, some long, some rapidly, some slowly rising or falling in pitch, some beginning and ending on acuter degrees of pitch, some on graver degrees, some passing from acute to grave, some from grave to acute. 1 Aristoxenus, harm. elem.% I, §28, p. 9, 20 Μ: την μέν ovv συνεχής λογικήν είναί φαμεν. διαλεγομένων γάρ ημών, όντως ή φωνή κινείται κατά τόπον, ώστε μηόαμον δοκεϊν ιστασθαι. κατά γε την ετέραν) ην ονομάζομεν διαστηματικήν, έναντίως πέφυκε γίγνεσθαι, αλλά γάρ ϊστασθαί τε δοκει, καϊ πάντες τον τοντο φαινόμενον ποιεϊν ονκ,έτι λέγειν φασίν, αλλ' αδειν όιόπερ εν τφ διαλέγεσθαι φευγομεν το έστάναι (ιστάναι libb.) την φωνήν, άν μη διά πάθος ποτέ εις τοιαυτην κίνησιν άναγκασθώμεν έλθεϊν έν δέ τω μελωδεϊν τουναντίον ποιονμεν. το μέν γάρ συνεχές φευγομεν, τό δέ έστάναι την φωνήν ως μάλιστα διώκομεν ' δσω γάρ μάλλον έκάστην των φωνών μίαν τε και έστηκνϊαν καϊ τήν αυτήν ποιήσομεν, τοσουτω φαίνεται τή αϊσθήσει το μέλος ακρι- βέστερον. οτι μέν δυο κινήσεων οϋσών κατά τόπον της φωνής, ή μέν συνεχής λογική τις έστιν, ή δέ διαστηματική μελωδική, σχεδόν δήλον έκ των εϊρημένων.ACCENT AND ACCENTUAL ARSIS AND THESIS. 63 It is not therefore surprising to find that the various systems of denoting accents by written signs agree in this, that they ignore certain kinds of glides and speak of acute or grave or middle tones, without further indication of their nature. It is clear that the purpose of this apparent defect in the notations, is only to simplify the theory. Even upon syllables of the shortest duration there can have been no perceptible fixity of intonation, such as is heard in singing. Aristoxenus and other writers recognize this point. The moment a tendency towards fixed intonations can be detected, the conversational manner ceases and singing begins. Continuous motion is abandoned for the intervallar. But, inas- much as the nature of a glide—its direction and extent—becomes more difficult to analyze in proportion as its duration is short, nothing essential is lost by marking short syllables or short vowels with only a general indication of the region of pitch in which they exist. But in the system of accentuation which ultimately prevailed, acute accents are found not only on short but also on long vowels, and it cannot be claimed that the glides on such long vowels were imperceptible or unimportant. In this case it would seem that the accent denotes an upward glide.1 The downward glide retained a special mark of its own, the circumflex accent. At one time it would seem that all syllables were marked with accents, but in course of time only those syllables in general which contained an acute element were so marked. This acute element was denoted either by the acute or by the circumflex accent sign. Every word, not enclitic nor proclitic, bore one such point of acuteness and one only. This doctrine is found in both Greek and Latin theory.2 1Brugmann, Griech. Gram.3 in Mtiller's Handbuch, §144, p. 151. 2Dionysius Hal., de comp. verb., xi : ταϊς δε πολνσνΤίλάβοις, όίαί ποτ1 αν ώσιν, ή τον οξύν τόνον έχουσα ιιία εν πολλαϊς βαρείαις ενεστιν. Cicero, or., XVIII, 58 : Ipsa enim natura quasi modularetur hominum orationem in omni verbo posuit acutam vocem nec una plus nec a postrema syllaba ultra tertiam. Quintilian, inst. or., I, 5, 30: namque in omni voce acuta intra numerum trium syllabarum continetur, sive eae sunt in verbo solae sive ultimae et in iis aut proxima extremae aut ab ea tertia. trium porro de quibus loquor, media longa aut acuta aut flexa erit, eodem loco brevis utique gravem habebit sonum ideoque positam ante se id est ab ultima tertiam acuet. est autem in omni voce utique acuta, sed numquam plus una nec umquam64 C. W. L. JOHNSON. Thus the melodic outline of a Greek sentence, and of a Latin sentence also, if the accentus of the grammarians was really the same as the Greek προσωδία, comprised a succession of summits corresponding to the accented syllables of the more important words. There was a periodic fluctuation in the tone from regions of low pitch to summits of high pitch. There was a rhythmical rise and fall, running through the sentence. In music proper the upward and downward movements, in which melody largely consists, received considerable attention, to judge from the somewhat complicated terminology which we find in the musical treatises. The usual words employed were βπίτασις and ανζσις. Bacchius, introd., §19, p. 6 M, p. 297 KvJ I MeXos δε τί Ιστιν ; —" Ανε,σις κα\ έπίτασις δι εμμελών φθόγγων γινομένη. Ibid., §45? ρ. 12 Μ, ρ. 302 KvJ : Πάθη δε της μελωδίας πόσα λεγομ,ε ν ςιναι \—δ'.—Τ ίνα ταύτα]—"Α,νζσιν, επίτασιν, μονη ν, στάσιν. "Ανεσις τί ε'στι ',—Κίνησις μεΧων άπο τον όξντερον φθόγγον €πι το βαρύτ€ρον. 'Έπίτασις δε τί Ιστιν \— Έπιτασίς εστί κίνησις μελών άπο τον βαρντερον φθόγγον €π\ το δεύτερον. Μονη δε τί εστίν \—"Οταν επί τον αντον φθόγγον πλείονες λέξεις μελωδωνται. Στάσις δε τι εστί ',—Στάσις εστίν νπαρξις εμμελονς φθόγγον. Gaudentius, introd., 1, p. 3 Μ, ρ. 328 KvJ : η δε της φωνής κίνησις εκ βαρντερον μεν εις οξντερον Ιονσης τόπον ειτίτασις, άνάπαλιν δε ανεσις καλείται re κα\ εστίν. Aristides Quintilianus, de musica, I, ν, p. 8 Μ, p. 5, 28 J : ταύτης1 δε είδη δυο, ανεσίς τε καϊ επίτασις* ανεσις μεν ουν εστίν ηνίκα ultima, ideoque in disyllabis prior; praeterea numquam in eadem flexa et acuta, quia in omni flexa est acuta, itaque neutra cludet vocera latinam. ea vero quae sunt syllabae unius, erunt acuta aut flexa, ne sit aliqua vox sine acuta. Servius, comm. in Don., p. 426, 15 Κ: unus autem sermo unum accipit accentum vel acutum vel circumflexum, utrumque autem simul habere non potest. Martianus Capella, III, p. 65, 22 Eyss.: omnis igitur vox latina simplex sive composita habet unum sonum aut acutum aut circumflexum. duos autem acutos aut inflexos habere numquam potest, gravis vero saepe. 1 It is not clear from the text to what ταντης refers. It cannot refer to τάσις of the preceding line. Perhaps it goes with της κατά την φωνήν κινήσεως four lines above.ACCENT AND ACCENTUAL ARSIS AND THESIS. 65 αν από όξντε'ρον τόπον εϊς βαρύτερο ν ή φωνή χωρη, επίτασις δ' όταν εκ βαρντερον μεταβαίντ] προς όξντερον. Plethon, Notices et ExtraUs etc., XVI, 2, p. 234: φωνής ανεσις [εστίν ή] επί το βαρύτερον μεταβολή, έπίτασis δε η επϊ το δξντερον, στάσις δε ή iv τω αντω οσαγε κατά την βαρύτητα η οξύτητα της φωνής μονή. But a more complete analysis of the melodic movements is found in a number of treatises. According to Aristides Quintil- ianUS μελοποιία has three forms, άγωγή, πεττεία, and πλοκή. The first of these is not defined, but it is divided into three varieties, which are named and described. 'Αγωγή ενθεΐα is an ascent by consecutive notes, άγωγή άνακάμπτονσα a descent of the same sort, while άγωγή περιφερής is a kind of combination of the first two, either ascending by the conjunct notes and descending by the disjunct notes, or vice versd. Arist. Quin., de mus., I, xii. p. 29 M, p. 19 J : αγωγής μεν ονν εΐδη τρία, ενθεΐα, άνακάμπτονσα, περιφερής* ενθεΐα μεν ονν εστίν ή δια των εξής φθόγγων τήν επίτασιν ποιονμενη, άνακάμπτονσα δε ή δια των επόμενων άποτελονσα τήν βαρύτητα, περιφερής δε ή κατά σννημμενων μεν επιτεινονσα, κατά διεζενγμενων δ άνιεΐσα, ή εναντίως ' αντη δε κάν ταΐς μεταβολαΐς θεωρείται. Bryennius, p. 5°2 Wallis, has the same analysis of άγωγή into ενθεΐα, άνακάμπτονσα, and περιφερής, but I have not access to a text of his treatise. The doctrine seems to go back to Aristoxenus, for we have a corrupt passage giving a similar definition. Aristox., harm. elem., II, §70 f, p. 29, 31 M, p. 28 W: 'Αγωγή δ' εστω ή δια των εξής φθόγγων εξωθεν των άρχων α)ϊ> εν εκατερωθεν άσύνθετον κείται διάστημα . . . ενθεΐα δ* ή επ\ το αντο . . · The same definition of άγωγή appears in Cleonides (Pseudo-Euclid), introd., p. 22 M, p. 207 KvJ : δι ων δε μελοποιία επιτελείται δ' εστίν' άγωγή, πλοκή, πεττεια, τονή. άγωγή μεν ονν εστίν ή δια των εξής φθόγγων όδος τον μελονς. 'Αγωγή is thus an Ιπίτασις or ανεσις of consecutive notes in the scale. A different set of terms, outlining a slightly different conception, is found in the anonymous treatise edited by Bellermann, Berlin, 1841, and by A. J. H. Vincent, Notices et Extraits des Manu- scrits de la Bibliotheque du Roi, Paris, XVI (1847) pt. 2, p. 5 ff. Anonymus, de musica, §16, p. 52 ff. (Bell. p. 19, nos. 2 ff. and 84 ff·) · ττρόσληψίς εστίν εκ τον βαρντερον φθόγγον επι τον66 C. IV. L. JOHNSON. όξύτερον κατ α μέλος επίτασις ήτοι άνάδοσις, ην τίνες κάλουσιν νφεν εσωθεν. τοντο δε γίνεται ποικίλως, αμέσως τε και διάμεσου' αμέσως μεν εκ τον εγγύς φθόγγου, οιον* Ε-1 Γ, TL, LF, FG, GO, ΟΠ, Π<[. εμμεσως δε οίον δια τριών F Ο , δια τεσσάρων FII, δια πεντε F < . εκληψις δε τα υπ εναντία τούτοις, άπο των οξύτερων επι τα. βαρέα ίίνεσις, ην τίνες δνομάζουσι νφεν εξωθεν, οίον αμέσως μεν GF, εμμεσως δε δια τριών Ο F, δια τεσσάρων IIF, δια πεντε <^F. πρόσκρονσις μεν εστίν εν χρόνοις δυο ενός, τουτ1 εστίν ελάττονοςγ χρόνου δυο μελη, τουτ εστί δύο φθόγγοι, άπο των βαρέων επί τα όξέα} οίον άμεσως μεν εκ τον εγγνς φθόγγον FG, εμμεσως δε δια τριών F Ο , δια τεσσάρων FIT, δια πεντε F