MATTEO BÀRTOLI (1873-1946)

Matteo Bàrtoli

Giacomo Devoto

TRANSLATOR’S NOTE: The text is a translation from the Italian, made with the help of my students. While we have striven to preserve the meaning and spirit of the original, we feel it necessary to warn the reader against the usual dangers of misunderstandings that may creep into translations. G. Bonfante.

1. Thirsting to know what our American colleagues have written and said in the years of separation, we often wonder, far away and poor, envious of their ample means for study, whether we are debtors only or whether occasionally we are able to contribute a worthwhile thought or idea. An occasion for this examination is afforded us by the anniversary of the death of Matteo Bàrtoli,¹ a talented and fascinating scholar, well-known outside of Italy also for his warm humanity and for his active participation in the work of international congresses.

Matteo Bàrtoli has been judged in various ways, and in America too, on the occasion of his death, conflicting opinions were read or heard. May an Italian linguist, a friend though not a follower of Bàrtoli, be permitted to attempt an outline of his character and of his place in the history of our science.

We have a curious impression when we reflect on the rich and valuable linguistic activity in America : that we are terribly young and open-minded. When we reflect on such characteristic works of American linguistics as R. G. Kent’s Sounds of Latin (1932), C. H. Grandgent’s Introduction to Vulgar Latin (1907), and E. H. Sturtevant’s Comparative Grammar of the Hittite Language (1933), we are struck, not only by the deep interest of their formulations, and of their results, but also by a calmness and, as it were, by a fixity of theory, as if between them and us there stood the difference between the adult who has now accepted social conventions as definitive and needs rather to impose them around and outside of himself like something to surround him and by that very fact to set him at rest, and the youth who discusses the truths of his fathers and wants to affirm in words and acts that nothing can be withdrawn from his free examination. And, in fact, the Sounds of Latin does not consider the problem of the Italic genealogical relationship ; Grandgent’s Vulgar Latin does not discuss the traditional relation between written and spoken Latin ; Sturtevant, at the very moment that he places Hittite in a position of its own with respect to the other Indo-European languages, reaffirms, amplifies and thereby precisely accepts the family tree theory of genealogical descent as a definitive conquest.

Certainly the effort of American linguists to reach conclusions makes discussions on method sometimes appear to them superfluous and distracting. But it must not be forgotten that, along with practical discussions of the method to be followed in each field of research, those discussions must continue which consciously examine the method of one’s work. The scientist must not only discover new concatenations, but he must also be able to say why, as a man aware of what he is doing and not as one governed by instinct or as a submissive and disciplined technician, certain concatenations have been affirmed, denied or replaced. It is a common belief that in science method comes first and then discovery. In reality discoveries, although presupposing a certain preliminary aptitude in reasoning and by that very fact a philosophy, are the fruit of an intuition or of an evident truth, and only at a later time can they be justified in a rational way. The observation of the similarity between pater and pita was at first intuitive, and only afterwards justified by schemes of regularized phonetic correspondences.

Intuition is in its turn preceded and guided by a mental aptitude in classifying relationships, in re-grouping and distinguishing, in coordinating and contrasting. An immense vicious circle represents the history of scientific thought, an effort to synthesize and regularize the infinite variety of the changes of nature by means of ‘ laws ‘, an effort to pursue in the smallest details the true realities in their variety and concreteness. The characteristic figure of a scientist can be delineated over and beyond the exterior balance-sheet of what he has done and what he has formulated, in an appraisal of his contribution both to refinement and simplification : progress must avoid the dangers, in the first case, of over-subtlety, in the second of over-simplification.

Even within the compass of the traditional comparative method, it has been possible to prove that its apparent simplicity conceals an interior complexity which does not lessen its value but which should be brought to the attention of active linguists. The excellent article of Giuliano Bonfante, On Reconstruction and Linguistic Method,² makes any demonstration of mine unnecessary. What should remain quite clear is that if in the traditional comparative method the spatial norm of the greater area (area maggiore) was already widely applied from the very beginnings of our science, this does not detract from the merit of Bàrtoli, its felicitous formulator. It is therefore not fair to say that in his formulas there is at times something not new and therefore not meritorious : he has made the researcher aware of a procedure which he formerly followed unconsciously ; his merit is the same as that of the first discoverer.

Born at Albona in Istria in 1873 ; professor at Turin from 1908 until his death ; a Romanist deeply concerned with problems of method from 1900, the date of his first work, to 1925 ; ardent theorist and successful user of the principle of the areal norms even outside of the Romance field from 1925 until his death, he became famous with his monumental work on Dalmatian,³ the Neo-Latin language of Pre-Venetic Dalmatia, exhaustively studied by him in its surviving traces, and brilliantly connected with the Italian dialects of the opposite Adriatic shore. He turned again to Dalmatian in his last days, with a nostalgic feeling for the works of his youth, tormented by the events which tore him from his native Istria : but unfortunately the revised edition of this monument of his will not appear.

Still, even if this essential document of his fervid activity is lacking, it is not difficult to form an idea of its spiritual development and of its ability to instruct through the two companion works Introduzione alla neolinguistica (1925) and Saggi di linguistica spaziale (1945). The systematic character of the first is completed in the second with the life blood of 129 publications, enriched with bibliographical citations and complete indexes.

2. In the classic manual of Indo-European comparative grammar, Brugmann’s Grundriss, the conclusions from the immense amount of material are presented, with regard to Latin, thus : ⁴ Latin is a language belonging to the Italic group, that is, it is a further offshoot of that branch of Indo-European languages defined by the characteristics en and em from syllabic nasals, or and 0l from syllabic liquids, kl from *tl, f from initial dh, z from intervocalic s, ablative singular in -ād, -ēd, -īd on the model of - ﾔd, instrumental plural in -ais on the model of -ois, ls. of the substantive verb *som instead of *esmi.

We have, thus, on the one hand a genealogical definition entirely rigid and linear, on the other, sporadic examples alined with no discrimination ; in the background, a generic state of mind, the impression that Latin, like other Western languages, represents a grammatical whole remarkably evolved and simplified, by comparison with Greek and Sanskrit, both from the point of view of vocalic alternations and of temporal verb forms.

The basis in R. G. Kent’s manual, The Sounds of Latin, is no different. Everything Latin in it is placed on a straight line, as on the finish-line of a race, and these facts are compared with an intermediary point, Italic, and this in its turn with the starting-point, Indo-European.

Bàrtoli’s article of 1934, II posto che spetta al latino nella famiglia dei linguaggi ario-europei ⁵ (his masterpiece) breaks up this rigidity and this uniformity and brings into the evaluation of these problems the breath of life with its complications. In place of the classic scheme he alines a series of elements from whose variegated sum rises the picture, different to the eyes of each person, of the ‘ position to be attributed to Latin.’ There are facts for which a correspondence can be found only in Sanskrit and Tocharian, or in any case in eastern regions of the Indo-European territory : they can be considered as archaic survivals whose continuity has been interrupted by intermediary innovations : ignis no longer in contact with agnis, endings in r no longer found in Greek but remaining in Tocharian. And so the foundation crumbles of the distinguishing criterion kentumIsatam which Terracini had already dismantled in 1921.⁶

Then there are Latin features limited instead to the Western languages, which in their turn can be interpreted in different ways, but always as archaic characteristics, according to the criterion of the ‘ greater area ‘ by which Latin nix lacks a correspondence in Sanskrit but is so widespread in all the rest of the Indo-European world that its antiquity cannot be doubted ; or according to the criterion of the more isolated area by which the meaning of sitis (‘ thirst ‘) compared with Greek phthisis can be considered the original one, assuming that Latin reflects an environment in which the idea of ‘ thirst ‘ did not undergo modifications with respect to the original Indo-European world, while in the more Eastern languages the aridity of the steppes must have exposed, although in a much greater area, the old meaning to modifications. Without taking a definite position in the polemic carried on by others against the intermediary ItaloCeltic and Italic unities, this method, which implies the passage of ‘ postethnic ‘ (see below) innovations from language to language, smashes the system of traditional genealogies—an affirmation to be compared with the still more decisive one concerning western Germanic, also opposed as a conservative zone to eastern Germanic,⁷ with an evident challenge to the genealogical concept of Proto- Germanic.

To this distinction of criteria which allow the opposition of conservative to new features, Bàrtoli’s method adds two further merits. The first consists in the search not only for the home (and consequently for the age), but also for the cause of the innovations, that is, for the environment which has permitted or imposed them. The expansion of the Indo-European languages brought with it the hypothesis of Indo-European migrations, which some, like Patroni, have wanted to underestimate by claiming that it was a question of ‘ few and barbarians ‘, but which Bàrtoli instead would make proportionate to the linguistic effects by imagining that they were ‘ barbarians but numerous ‘, and for which perhaps the best solution is found in the formula ‘ few but compact ‘, which alone can explain both the importance of the linguistic revolution and the absence of cultural repercussions. The second merit of the work consists in the fact that the various so-called areal norms are presented not as necessary but as statistical laws, affirming that ‘ usually ‘ the examples correspond or that the majority of the examples is positive and the minority negative.

In order to arrive at a final equilibrium, we should however be able to weigh the examples which are based on the isolated area or on the greater area, since number alone cannot allow us to assert that the conservative features predominate over the innovations. The importance of the loss of final -s or of the sonorization of intervocalic consonants, which separates Gaul from Italy by a distinction rich in historical and cultural consequences, is one thing ; the difference between fiume from flūmen and fleuve from flŭuĭŭs, which has very little importance, is quite another.

It would be valuable to have a recasting of the Sounds of Latin which would take into account the concepts of Bàrtoli’s work and which would graft the Italian feeling for history to the philological solidity of the American work.

3. Bàrtoli set out on the road to perfecting the essay on Latin in the two works Per la storia del latino volgare and Per la storia della lingua d’Italia, which appeared in 1927. He drew the conclusions shortly afterward (in 1936-37) with the Caratteri delle lingue neolatine.⁸ Here too the areal norms constitute the decisive novelty, especially that of the lateral areas already mentioned, by which the old rogâre survives in Spain and Romania but is replaced by precāre in France and Italy ; and that of the later area which preserves archaic elements introduced with the most recent conquest, whereas in the original area of Italy these elements have been exposed to innovations and replaced : thus the innovation chièdere in the earlier, original Italian area compared with the archaic type Spanish pedir preserved in the later area ; not to mention the norm of the isolated area which explains how Sardinia (and especially its central region Logudoro) is a stronghold of valuable survivals such as the pronunciation ke, ki of Lat. ce, ci.

Here too, along with the geographic technique, the historical consideration springs alive through the examination of the causes of the innovations. And in the face of mechanical tendencies, so often revived, toward indiscriminate Pre-Roman influences supposed to react, altering and deforming the linguistic unity of Latin, Bàrtoli prefers to stop and consider reasons innate in the Latin world, the administrative order of the empire, the reform of Diocletian, the prestige by which the culture of Transalpine Gaul, rather than the famous remains of the original and, so to say, physiological Gallicity, would be the basis of the Gallo-Latinity of Cisalpine Gaul. On the other hand, because of the importance which must be attributed to imitation in linguistic matters, the substantial affinity must have favored the influence, generally underestimated, of the Osco-Umbrian linguistic tradition over the weak tradition of the Roman literary language, and it is to Bàrtoli’s credit that he recognizes it explicitly.⁹

In America also, at almost the same time, studies in Vulgar Latin received a vital impulse with H. F. Muller’s Chronology of Vulgar Latin, a stirring book, the importance of which will never be sufficiently emphasized. But the reaction which this book represents in comparison with traditional concepts is essentially of a quantitative order, it cannot be an end in itself : it represents the thesis of the opposing party, in favor of the continuity of the tradition of the literary language, expounded coherently, with compelling arguments. For this reason it is called legitimately a ‘ chronology ‘ and not a history. For this reason, it is to be hoped that there may some day be a recasting of this work which, without changing the general theses, takes into account that life and that moment which is inseparable from the life of the Latin language in the heart of the imperial age, and which Bàrtoli as a pioneer has brought to light. The tradition of the Latin literary language can have been solid and firm, and at the same time furrowed by regionalisms more and more accentuated, which can very well have overcome resistance only with the Carolingian period ; but not on this account, as Bàrtoli has shown, must they be judged non-existent in the preceding centuries.

4. Bàrtoli’s central formula is that ‘ Every word has its own history and every phase its own area.’¹⁰ From this it is clear that the so-called ‘ areal norms ‘ set forth in the fundamental Introduzione alla neolinguistica (1925) are not really ‘ norms ‘ or ‘ laws ‘ according to the current meaning of the word, but only type-possibilities among which the researcher must always know how to choose with a sense of responsibility. One rises from the fact to a comfortable framework within the patterns of the norms, and not vice-versa : from the abstract norms one does not descend to do violence to the facts. Thus Bàrtoli fulfills his first task, that of bringing into 19th century linguistics (diachronie since it confronted two phases distant in time, but not yet historically concrete) the concreteness, the singularity of history ; thus Bàrtoli fulfills in his first phase that indispensable task of the scientist which consists of making things particular and concrete, of refining and complicating.

But the work of ‘ complication ‘ is only a moment in the eternal development of thought which analyzes, divides and then regroups and distinguishes according to new standards. The simplifying action practised in the field of method by the formulation of the areal norms is thus accompanied by the excellent distinction, in the field of Indo- European linguistics, between ‘ pre-ethnic ‘ innovations and ‘ postethnic ‘ innovations, and in Romance linguistics between innovations of the ‘ Roman period ‘ and innovations of the ‘ Romance period‘ .

The first distinction goes back to 1934 ¹¹ and is substituted in the traditional picture for the distinction between genealogical innovadons (in Schleicher’s sense) and casual coincidences. The innovation which distinguished the type ｺ from the type к is a pre-ethnic innovation, which precedes the splitting-up of the Indo-European languages. The innovation which strongly labialized the labio-velars in East Celtic, Umbrian and Greek is a post-ethnic innovation, which in the distinctions by now perceptible between the various Indo- European languages did indeed find barriers, but contiguous and surmountable ones, and not a void to be overcome by ‘ leaping ‘ from one branch to the other.

In the Romance branch, innovations like those which weakened and finally eliminated the final -s or affected the pronunciation of the groups ke and ki are innovations of the Roman period (due moreover to pre-Roman influences, Osean, Umbrian) which could not reach all the frontiers of the linguistic territory of Latin. On the other hand the diphthongization, though it may have some presupposed Roman element in the lengthening of the accented vowels, developed in the Romance period although characters not corresponding to the limits of the future individual Romance languages; i.e., it presupposes exchanges and penetrations at a time in which the characteristic and definitive innovations had already affirmed themselves.

The contrast with theses like that of Indo-Hittite unity or of the homogeneity of pre-Carolingian Latin could not be sharper. At the same time anyone can see how opportune it is to have the same facts and the same arguments of the two American scholars taken up, completed, contradicted or confirmed through Bàrtoli’s concepts of the areal norms, of the pre-ethnic—post-ethnic and Roman— Romance strata.

5. The scientist is human. In the eternal flux of distinguishing and regrouping, of analysing and summarizing, there are times when, caught by the current and born along by his very weapons, he wants to break the closed circle and, going beyond simplification, to construct ; beyond subtilization, to attenuate, to dissolve.

From the dissolution to which insistence on the individuality of the changes in words would have led him, Bàrtoli saved himself by reacting instinctively against the honest but hybrid efforts—’ glottosophical’ he once called them—of Vossler and his school. But not from reconstruction. To reconstruct, statistical laws such as are realized in the areal norms are insufficient ; exacting criteria are necessary, that will have the same force of conviction as logicomathematical ones. It is the same psychological attitude as that of a man who, equipped with weapons of defence, feels rise within himself an ever clearer awareness of his privileged strength, and little by little becomes dangerous, in the same way that states through the possession of arms become imperialistic.

By an almost inevitable sequence this suppression of criticism, this psychological attitude first appears in Bàrtoli the very year (1925) in which, by formulating the principles of ‘ neolinguistics he had shown that he had become fully aware of the instruments of his work. It can be seen in a short article which appeared in a provincial journal, the Bollettino della società filologica friulana, with the title, Di una legge affine alla legge Verner. In this article, pregnant with consequences, he maintained the thesis that the Indo-European voiced aspirate stops were not primitive, but arose (by a process similar to that of Verner’s law) from a particular position of the simple voiced stops with respect to the accent.

The spectacular deductions to which he remained faithful, in spite of the opposition and total dissent of scholars, becomes more and more rigid, and the clearest examples are precisely from the last five years, 1940-45.¹²

The conclusions drawn from these constructions can be summarized in 11 postulates, rigidly bound together not by a logical chain of reasoning, but by the empirical necessity of avoiding the internal contradictions which kept multiplying. Of these postulates, that of ‘ Indo-European oxytony ’ is of capital importance. It brings with it the necessity of affirming that ‘ all final stops were voiced ‘ ; that aspiration arose only when they were followed by an accented vowel ; that when we find posttonic (and not immediately protonic) aspirates, a shift of accent must have intervened (méthy from *methii, élaphos from *elabhós, thymós from *dhŭmos) ; that when we find final voiceless stops we are confronted with stems originally ending in a vowel ; that according as the consonant is voiceless or voiced, we can think of old nominal stems ending in vowel or in consonant respectively, or also of double stems of the type *gâd, the basis of Gothic gôds, and *gados, the basis of Greek agathós ; finally, that the neuter and feminine grammatical genders prefer monosyllabic forms ending in a consonant.

Of these postulates, which by the very nature of the ‘ postulate‘ are indemonstrable, none can be accepted. Bàrtoli’s justification lies only in the fact that in reality, in the traditional representation of the Indo-European roots, the difficulties of the oscillations of the type *bheug(h)- have only a graphic solution and leave open a problem which we are wrongly accustomed to skipping over.

Bàrtoli goes back to Benveniste and his well known work on L’origine de la formation des mots,¹³ showing points of agreement and disagreement. But the difference in inspiration between the two works is profound, even though the linguist endowed with a feeling for history may actually find affinities. Benveniste’s urge is of a systematic nature ; he is conscious of exercising ‘ from without ‘ an effort of a geometrico-mathematical order to represent rationally what is apparently irregular and complex. It is a Saussurian offshoot and one can discuss whether it is intrinsically a happy one, but it is not open to criticism unless it has pretensions of historical order. Benveniste can boast of no new skill in method, of no new hermeneutic key.

Bàrtoli’s urge, on the other hand, is to arrive at a historical reality surprisingly different from the one expected and commonly accepted, thanks to the ‘ secret weapons ‘ which he thought he held at his disposal. These fail because the proposed objective was too lofty : weapons can only destroy, they do not construct. Therefore Bàrtoli’s postulates become as rigid and artificial as the criticized phonetic laws of the positivistic period.

And in this return to the positivistic 19th century lies the drama of Bàrtoli who in fact affirms in the conclusion of his last work, that in his life he sought not to deny laws, phonetic or otherwise, but only to י reform them ‘. He complacently alines in the index of the Saggi the lex Bartholomae along with the laws of Bàrtoli, Brugmann-Pisani, Hirt, Lachmann, Mussafia, Osthoff, Saussure-Fortunatov, Streitberg, Verner, and Walde.

Precisely while the physical sciences were turning towards the ‘ statistical ‘ notion of laws and now consider hypotheses as architectural constructions whose first merit must be simplicity, Bàrtoli was falling back into their orthodox veneration : his formulations were becoming mechanical, arbitrary, complex. For this reason, few scholars‘ lives are as instructive and admonitory as his. The development of his thought has, like the life of man, an organic evolution : a dawn, an apogee, a decline. It is, from the documentary point of view, a perfect example : it should be observed, meditated, and remembered like all experiments which could be pushed to their extreme conclusions.

Intent on going his way without rest, Bàrtoli sometimes showed himself less sensitive to the attractions of other methods of work. He was deeply curious about the most diverse linguistic families, so quick to apply the areal norms to the comparisons of pre-ethnic Indo- European with Finno-Ugric (and in general non-Indo-European) innovations ; he worked even outside of linguistics to make them known among ethnologists and folklorists ; except for Trombetti and Tagliavini, he was the only Italian scholar to concern himself with American languages.¹⁴ Linguistic paleontology barely touched him, when he thought he saw a trace of polygamy in the importance which plural forms supposedly had in the system of words for mother, son, compared with pater. Taste for phonemics, syntax, stylistics, general grammar, the description of a system in general, remained foreign to him.

Closed in the cycle of his life, he died while his native land, his own country stands at a decisive turning and all of Western culture is confronted with an examination of conscience. May both turning and examination signal a happy renewal in linguistic studies as well.

Source : Giacomo Devoto,י Matteo Bàrtoli,’ Word 3.208-216 (1947). By permission of Word, and the author.

¹ At Turin, January 20, 1946.

² Word 1.83 ff., 132 ff. (1945).

³ Das Dalmatische Altromanische Sprachreste von Veglia bis Ragusa und ihre Stellung in der apennino-balkanischen Romania, 2 vols., Vienna, 1906.

⁴ Vol. 1, p. 10.

⁵ Republished in Saggi di linguistica spaziale 1945.1-31.

⁶ In Atene e Roma of 1921.

⁷ Saggi, 213 ff.

⁸ Saggi, 75-119.

⁹ Saggi, 100 ff., 134 ff.

¹⁰ Saggi, 215.

¹¹ In the article cited in note 5.

¹² See in Saggi : La questione delle medie aspirante e la coppia greco agathós e gotico gods (buono) 180 ; Il ritmo dei tipi patér e métër e la poligamia arioeuropea 170 ; Monosillabi e bisillabi di età pre-etnica 243.

¹³ Paris, 1935.

¹⁴ Le origini degli Indiani d'America lumeggiate dalle aree linguistiche, Annuli dell'lstituto superiore di magistero di Torino 335-52 (1934) ; Ancora delle origini dei linguaggi precolombiani alla luce delle norme spaziali, Melanges Van Ginneken 123-33 (Paris, 1937). Cf. also the Atti del III Congresso internazionale dei linguisti, Rome 1933, 419-26 (published Florence 1935) ; Linguistics spaziale, in Biasutti, Le razze e i popoli della terra, 1, Turin 1941, 329-33 ; Scritti Trombetti 175 ff. (Milan, 1930).