Humboldt's Description of the Javanese Verb

[I conducted the research on which this paper was based while I was a staff member of the Mechanical Translation Project at the Massachusetts Institute of Technology (1960-1964). At that time, I was able to attend the lectures on linguistics and the history of linguistics given by Noam Chomsky in the Institute and was fascinated by his portrayal of Wilhelm von Humboldt as (in spirit at least) a "generative" grammarian. I particularly recall that Chomsky focused attention on Humboldt's concept of "linguistic form," never fully fleshed out in the famous introduction to his multi-volume treatise on Old Javanese and the other Malayo-Polynesian languages. For the text of that introduction, see Humboldt 1836 and Leitzmann 1903, the critical edition of Humboldt's writings commissioned by the Deutsche Akademie der Wissenschaften. At that time, it seemed to me that I had the ideal linguistic background to explore the question of how Humboldt applied that concept in his discussions of various languages in the "Introduction" in that I was writing a dissertation for Yale University on the grammar of one of the Austronesian languages (Toba-Batak) and had earlier taken two years of Sanskrit while I was a Linguistics graduate student at Yale. The Widener Library at nearby Harvard University, to which I fortunately had access then, owned copies of the magnificent three volumes of Humboldt's "Kawi-Spache auf der Insel Jawa." It therefore seemed the most natural thing in the world for me to consult the relevant volumes of that series and see how Humboldt described Old Javanese grammar. Perhaps in that way, I reasoned, we would discover how he applied the notion of linguistic form in a concrete instance. This paper, therefore, resulted from my grappling with the primary data cited by Humboldt himself. Let me say that the result of this line of investigation was inconclusive. To my surprise, it led to no new insight into Humboldt's concept of linguistic form. In this paper, however, the reader will notice that I make some general comments about Humboldt's linguistic procedure, which I found to be exemplary. Peter Schmitter, a well-known Humboldt specialist, once told me that he had found my paper valuable and regretted that I did not continue to work on Humboldt. I discovered later that research on Wilhelm von Humboldt tends to be extremely controversial, so I do not regret my decision to steer clear of that field for most of my career. Needless to say, it is regrettable that scholars do not make a better effort to tone down the bitterness that they feel towards each other. After all, the problems we face are difficult enough without our adding that admittedly human element to the equation.

Expositions of Humboldt's linguistic writings have concerned themselves exclusively with his theoretical papers and his treatise on the diversity of language structure.[1] The purpose of this paper is to focus attention on what is perhaps one of his less appreciated facets, namely his contribution to descriptive linguistics. To illustrate this aspect of his work I shall examine a morphological problem which Humboldt encountered in his study of Javanese. Let me recall the fact that the famous treatise on linguistic diversity was but the introduction to a three-volume study of the Old Javanese literary language and its position within the Malayo-Polynesian family of languages.[2]

Humboldt's aims in writing the work were complex. First of all he wished to refute the view, defended by a number of scholars at that time, that Old Javanese was related to Sanskrit in the same way as Pali was, or in other words that Old Javanese was a linear descendant of Sanskrit with an admixture of local (i.e., Malayan) elements. Eduard Buschmann's foreword to the second volume made this aspect of the work clear (Humboldt 1838: III ff.), and one can also compare Humboldt's own discussion of the views of his predecessors (1838: 188-203).

It is clear, however, that the book was intended to do more than correct contemporary misconceptions about the genetic affiliations of Old Javanese. Humboldt also wished to characterize the structure of the Malayo-Polynesian languages in a general way, so that he could compare them typologically with Chinese and the languages of the Sanskritic family. This was in turn related to his more general purpose of exhibiting the interrelations between the linguistic and cultural development of the human species.

A description of Javanese was necessary, therefore, both to support his claim that the Old Javanese literary language was basically a Malayan, not a Sanskritic language, and to fill out his general characterization of the structure common to all Malayo-Polynesian languages. However, to describe Javanese adequately was no easy task in the thirties of the nineteenth century. Humboldt meticulously consulted all available sources of information on the modem language.

These included a translation of the Holy Bible into modern Javanese by a German missionary, G. Bruckner, a grammatical sketch of modern Javanese together with a reader and word-list by another German missionary, J.F.C. Gericke, a Javanese word-list compiled by the British governor-general of Java, Sir Stamford Raffles, and a brief Javanese grammar by another Englishman, John Crawfurd. Humboldt also carried on a correspondence with a young Dutch scholar, Roorda van Eysinga, who had gone to Java and was studying the language on the spot. (Roorda van Eysinga was later to publish a grammar of Javanese, which Humboldt, however, did not live to see.)

For the old literary language, however, Humboldt performed the extraordinary feat of constructing the grammar himself on the basis of a text consisting of portions of an epic poem, the Brata Yuddha, which had been published by Sir Stamford Raffles. In addition to all this, as part of his more general project of characterizing the structure common to the Malayo-Polynesian languages, he examined the grammars of all related languages he could lay his hands on, in particular those of Malay, Tagalog (Philippines), and Malagasy (Madagascar). Humboldt emphasized (1838: 322) how especially helpful his knowledge of Tagalog was in analyzing Javanese.

Since he shared with many of his contemporaries the view that language relationship can be proved if a set of languages can be shown to possess the same grammatical structure, he felt that were he able to demonstrate that the grammatical structure of these languages (Javanese, Malay, Tagalog, and Malagasy) was basically identical, and to show what this basic structure was like, he would have achieved his twofold aim of vindicating his supposition as to the true nature of Old Javanese, and comparing the structure of Malayo-Polynesian with those of Chinese and the Sanskritic languages.

However, although Humboldt's study of Javanese grammar was merely one link in a long chain of demonstration, it was none the less an essential link and one which he did not treat lightly. To show how conscientious his procedure was I shall examine a rather lengthy argument which he presented as part of his analysis of the Javanese verb.

The problem he faced was the following. Many Javanese verbs appear to be represented by a pair of stems which differ from one another in the initial consonant.[3] Thus the Javanese verb "write" is represented by the pair of stems nulis and tulis. Humboldt found the most comprehensive description of this phenomenon in a grammatical sketch of Javanese published in 1831 by the German missionary, J. F. C. Gericke. Gericke showed that in certain cases the shift in the initial consonant is a device to derive substantives from verbs. The following are some of Gericke's examples, as quoted by Humboldt (1838: 88-89):

marêntah	"to order"	parêntah	"order"
miyos	"to be born"	wiyos	"birth"
neda	"to eat"	teda	"food"
nyenjata	"to shoot"	senjata	"gun"
nyatur	"to relate"	chatur	"story"
ngaranni	"to name"	haran	"name"
ngrasa	"to feel"	rasa	"feeling"
ngililanni	"to concede"	lila	"concession"

Thus for Gericke one of the functions of the shift was to derive a noun (e.g., teda "food") from a verb (neda "to eat"). The two sets of alternating consonants according to Gericke were as follows:

First set	Second Set
m	w or p
n	t
ny	ch or s
ng	h
ngr	r
ngl	l

To make the ensuing discussion clearer I shall refer to forms with initial consonants from the first set as type A stems, and forms with initial consonants from the second set as type B stems. Thus in the case of the pair neda "eat" — teda "food" I shall call the former a type A stem, and the latter a type B stem.

Gericke, however, did not regard the difference between type A and type B stems as coinciding with the division of words into substantives and verbs. Type B stems, besides appearing in many cases as derived nouns, also occurred in the following forms:[4]

1. In passive forms.

2. In the imperative form.

3. In substantives derived by affixation.

4. In certain types of derived verbs.

Examples will make these facts clearer. Thus the type B stem tingngalli "see" occurs in the following forms:

1. In the passive form dhipun-tingngalli "seen."

2. In the imperative form; for example, in a sentence such as sampêyan tingngalli "Look!" (sampêyan "you").

3. In the derived substantive tingngallan "sight."

4. In the derived verb tingngal-tinningngallan "to look at one another."

Finally Gericke also observed that in some verbs the type A stem appeared preceded by the prefix ha- with no apparent change of meaning. Thus in addition to the form marêntah "to order" was the form hamarêntah with the same meaning.[5] Gericke called such forms lengthened forms of the verb.

To sum up, a fully conjugated verb, according to Gericke, reveals two types of stems: the type A stem, e.g., nulis "to write," and the type B stem, e.g., tulis appearing, for instance, in the passive dhi-tulis, and functioning as a derived substantive (tulis "piece of writing, letter"). All verbal and nominal derivatives formed by affixation contain the type B stem. A verb with a prefixed ha- is a lengthened form of the verb, or as Humboldt expressed it, a "fortuitous lengthening of the word." [6]

Having expounded Gericke's description in some detail Humboldt went on to present his own interpretation of the facts. But before doing so, he emphasized that he regarded Gericke's description as adequately reflecting contemporary Javanese usage, and that what he was about to say was in no way a contradiction of what Gericke had stated. For his own purpose in investigating these phenomena was a different one from that of his predecessor. Gericke had compiled the rules of Javanese to enable a reader so desirous to learn to speak and understand the language, whereas his own aim was to discover the deeper significance of the grammatical forms of Javanese, their original interrelations, and the process by which they came into being, often changing their character radically in so doing. This enterprise was fraught with uncertainty but could not be shrugged off, since it was the only way of discovering the structure of the language, and the nature of its relationship to other languages of the same family. But a prerequisite to any progress in this type of endeavor was strict reliance on the facts as revealed by those who have learned to speak the language. Their testimony must be followed without question.[7]

After stating his general aim, Humboldt then proceeded to reinterpret Gericke's description of Javanese verb morphology and in so doing completely altered Gericke's picture of the facts. The issue which Humboldt took up was precisely the relationship between type A and type B stems. Gericke, let us recall, had claimed that the type B stem was derived from the type A stem by a shift in the initial consonant.

It was on this initial consonant alternation that Humboldt focused attention. He once more posed the question as to which of the two sets of alternating consonants was the derived one, and unlike Gericke concluded that the first set was the derived set. He argued as follows. The initial consonants of the first set are sounds belonging to a single class, namely the class of nasals. The initial consonants of the second set, on the other hand, belong to no particular class. But the mere fact that the consonants of the first set share a common property indicates that they are derived, since the possession of a common feature presupposes a common cause, the operation of some special process. Conversely, the fact that the second set of sounds shares no property stamps it as original and non-derived, since it is precisely this state of affairs that characterizes the initial sounds of a language. That is to say, no language restricts the initial sounds of its non-derived words to one phonetic class. Any set of alternants whose initial sounds belong to only one class must therefore be derived, not original.[8]

Humboldt then showed that many words with initial consonants belonging to the second set were indeed original words, and more specifically original nouns, from which verbs were then formed at a later date. Thus the Sanskrit viçesa "power" was borrowed into Javanese and appears as the noun wisesa from which was formed the verb misesa "to have power," by the shift of w to m. Thus the historical facts corroborate what Humboldt had inferred on general grounds.

But he was not content to leave the matter there. Having shown which of the two sets of initials is the original one and which the derived one, he then proceeded to explain the consonant shift itself. His arguments involved two considerations: the articulatory nature of the shift in Javanese, and the conclusions that can be drawn from analogous consonant shifts in related languages. As to the first point, the shift itself replaces the initial consonant by its corresponding nasal. The comparative evidence is as follows: In Tagalog, Malagasy, and Malay, when a verbal prefix man- occurs in front of a stem beginning with a consonant, the latter is replaced by the corresponding nasal and the final -n of the prefix is dropped. Thus in Malay the stem tolong "help" appears after the prefix as me-nolong.

Humboldt then provided a step-by-step description of the shift.[9] He claimed that the resultant initial nasal of the stem in the Malay word menolong, for example, arose from the final -n of the prefix through loss of the original initial consonant of the stem, in this case t-. The final -n of the prefix remained unchanged before dental sounds but shifted in front of other sounds into the homorganic nasal (e.g., final -n becomes -m before initial p-), and then became the initial consonant of the verb stem. Schematically arranged a hypothetical stem such as pata might undergo the following changes:

Stage 1	man-pata
Stage 2	mam-pata
Stage 3	mam-ata
Stage 4	ma-mata

Humboldt then proposed that the same set of changes could be used to explain the Javanese facts, with the two minor differences that in Javanese the prefix has no initial m-, and that short forms exist without the vowel a. (Note that in the traditional orthography, which Humboldt used, an h was prefixed before an initial vowel, but was silent in the pronunciation, as Humboldt himself pointed out.) Thus in Javanese a hypothetical stem of the shape pata would have a double treatment:

	Long Form	Short Form
Stage 1	han-pata	n-pata
Stage 2	ham-pata	m-pata
Stage 3	ham-ata	m-ata
Stage 4	ha-mata	mata

But once again Humboldt was not content merely to describe the two developments, but went further and posited that the shorter form was derived from the longer form by a process of vowel reduction. The evidence for this proposal was of two kinds. First, he noted that in Javanese a trisyllabic stem with an initial vowel often drops that vowel when suffixes are added to it. Thus the stem hupama "similarity" has the form pama before the possessive suffix in the word pamane "its similarity." Thus a process of vowel reduction must be posited for Javanese in any case. It is then merely a question of extending its scope to include the examples we are concerned with.

His second argument was of a more general nature. Similar vowel reductions occur in almost all languages, particularly if one takes dialects of the common people into consideration, and here he cited German runter and rein from herunter and herein. Finally, he quoted attested doublets, in Javanese, that is, long and short forms of the same word, e.g., hamunduti and munduti from punduti (no gloss provided), and listed examples from Javanese itself of nasal stems occurring with a prefix ma-, e.g., ma-mukti "to eat" (derived from bukti "food," cf. Sanskrit bhukti).

Humboldt's account of the facts was thus as follows. In Javanese, as in other related Malayan languages, stems were nominal in character (unlike the predominantly verbal roots of the Sanskritic family of languages). Verbs were formed from these nominal stems by the addition of the prefix man-, which entailed a series of sound changes in its own final consonant and the initial consonant of the accompanying stem. Thus verb forms were clearly marked as such by the appearance of a prefix and the associated alternations, and were thus distinguished from corresponding nominal forms. In Javanese an additional development took place. First, the initial m- of the verbal prefix disappeared, and then the vowel a of the prefix was also lost. This left only the final n of the prefix, the only physical manifestation of which was a set of alternations between the original initial consonants of the stems and the corresponding nasal continuants.

Thus in Humboldt's view the difference between a nominal and the corresponding verbal form in modern Javanese rests entirely on the consonant alternations, as Gericke described them. The modern speaker of Javanese is, however, unaware of the true significance of the alternations, taking them to be merely a device to distinguish nouns from verbs.[10] That is, Humboldt accepted Gericke's description of the facts as a true reflection of the native speaker's conscious intuition concerning them. Thus in emphasizing the importance of accepting the observations of linguists familiar with a particular language, he seems to have been recommending that these observations be equated with the native speaker's knowledge of the language. There was perhaps also an implicit assumption that native speakers know very little more about their language than a linguist who writes an observationally accurate grammar adequate enough for the needs of those who want to learn the language. That is, it may be that Humboldt's view of the native speaker's knowledge of his own language was a rather restricted one.

In conclusion, certain general features of Humboldt's approach to description may be briefly pointed out. It seems clear that he had an understanding of the phonetic dimension of language. The notion of a class of sounds sharing a phonetic feature (in this case nasality) was a familiar one, as was the notion of a point of articulation. He also had definite, though unformulated notions about the distribution of phonetic features in base forms (stems).

In validating his inferences Humboldt aimed at explanations of as wide a generality as the facts warranted, and at no point did he disregard analogous facts from related languages or from languages in general. In this he differed from those descriptive linguists of the twentieth century who are inclined to suspect that an explanation is spurious if it appeals to facts outside the language under study.

In one other respect Humboldt's practice diverged from that of many present-day linguists. This has to do with his conviction that pure description is not a sufficient goal for the linguist. This means in practice that he constantly discussed diachronic phenomena while describing a language. This does not, however, mean that he confused description and history: he cited diachronic facts for their explanatory value. Thus Humboldt felt he had to establish the series of changes which culminate in a modern Javanese verb form in order to render such forms understandable, and such understanding was for him an indispensable goal of all linguistic work. But he did not claim that the derivation reflected conscious knowledge available to the native speaker.

Thus although Humboldt did not draw an explicit distinction between synchronic and diachronic aspects of language he leaned, at least in the case we are considering, toward the assumption that the complete set of shifts as he described them did not form part of the grammar of present-day Javanese, and that Gericke and other grammarians of the modern language were perfectly justified in describing them in a different way. I conclude, therefore, that the distinction between diachrony and synchrony, which was to loom so large in twentieth-century linguistic theory, is already implicit in Humboldt's work, and that for him synchronic grammar is equatable with that part of grammar of which the native speakers are consciously aware.

 

E N D N O T E S

1. For a bibliography of Humboldt criticism see Arens 1969: 763. Humboldt's significance as a theoretician of language has been discussed by Chomsky in a recent publication (1966: 19-31). It may be noted, however, that Chomsky's interpretation of Humboldt's theories has been challenged. See for example Zimmer 1968: 293-295. I am greatly indebted to Chomsky for directing my attention to Humboldt's work. I should also like to acknowledge much valuable stimulation from conversations with John Viertel. [For a discussion focusing on the significance of Humboldt's notion of "form" in relation to contemporary generative theory, see Chomsky 1964: 918-921. Go back to the text.]

2. Most of the quotations in this paper are from the second volume of the work, which appeared posthumously in 1838. For a recent description of Old Javanese, see Teselkin 1963. A description of modern Javanese can be found in Kiliaan 1919. Go back to the text.

3. The orthography which Humboldt used to transcribe Javanese forms has the following peculiarities: dh is the dental and d the retroflex stop, ny is the palatal nasal, and ng the velar nasal, ngng indicates a sequence consisting of a velar nasal followed by a velar stop, y is the palatal semi-vowel, and j and ch the voiced and voiceless palatal affricates respectively. Medial resonants often occur geminated but were not so pronounced. Go back to the text.

4. I omit mention of the syntactic environments of the type B stems which Humboldt discussed at some length. Go back to the text.

5. Gericke himself did not quote the form hamarêntah, but it is more suitable for purposes of exposition than the forms he cites which are all from stems with initial voiced stops, a category of stems whose behavior neither Gericke nor Humboldt were clear about. Go back to the text.

6. The original phrase was "eine gleichgültige Verlängerung des Wortes", the implication being that the lengthening was without function or meaning. Humboldt believed this to be a very significant feature of modem Javanese for in a summary characterization of the language later in the book he pointed out that certain grammatical features give Javanese the appearance of an inflectional language [den Schein, den flectirenden Sprachen näher zu stehn], but he continues: "Sie [die Javanische Sprache] entledigt sich dadurch der vielfachen, die Tag. Wörter beschwerenden Vorsylben, zerstört aber wieder durch eine gleichsam willkürliche Anhängung solcher verlängernden Laute diesen Eindruck ohne Nutzen, da diese Vorschläge der Bedeutung nichts hinzufügen" (Humboldt 1838: 321). The hidden assumptions here are that syllables or formatives which have no directly assignable meaning are a blemish on the language, and that what Gericke had called lengthened forms of the verb are indeed to be regarded as such from a synchronic point of view. Humboldt thus had very definite ideas about the ideal relationship between form and meaning, and it was against this ideal that he measured all the linguistic phenomena which he studied. Go back to the text.

7. The original passage reads: "Es ist ganz etwas anderes, wie er [Gericke] es zur Absicht hatte, die Regeln einer Sprache zum Behuf der Anwendung beim Verstehen und Sprechen derselben darzulegen, und, wie ich es hier versuche, der wahren und inneren Bedeutung der grammatischen Formen und ihrem ursprünglichen Zusammenhang, dem Wege ihrer Entstehung, auf welchem sie sich oft wesentlich verändern, nachzuforschen. Man stösst bei einem solchen Unternehmen oft auf Punkte, wo nur der Vermuthung Raum bleibt, darf es aber darum nicht aufgeben, da man nur auf diese Weise die Analogie der Sprache und ihr Verhältniss zu den stammverwandten zu erkennen vermag. Die erste und nothwendigste Bedingung zum Gelingen eines solchen Versuches ist aber die genaueste Beachtung des factisch in der Sprache Vorhandenen; und dies kann man nur bei denen finden, welche die Sprache unter den Eingebornen, aus langer Uebung und eigenem Gebrauche kennen gelernt haben; ihrem Zeugniss muss man sich daher in allem Factischen ohne Widerrede unterwerfen" (1838: 91-92). Go back to the text.

8. Humboldt expressed the matter thus: "Die Uebereinkunft der Laute deutet an, dass sie etwas Gemeinschaftliches an sich tragen, und weist dadurch auf eine gemeinschaftliche Ursach, eine künstliche Bildung hin, indess die zweite Reihe die Verschiedenheit der Consonanten an sich trägt, welche die Anfangsbuchstaben der Wörter einer Sprache überhaupt haben; in ihr scheint also der natürliche, ursprüngliche Zustand zu liegen" (1838: 97). Go back to the text.

9. Humboldt described the shifts in the following way: "Im Tag., Mad., und eigentlich Mal. geht, wenn das Verbalpräfix man vor die Consonanten der zweiten Reihe tritt, jeder von diesen in den entsprechenden Nasenlaut über, oder dieser Nasenlaut entsteht vielmehr aus dem End -nn des Präf. wird also, vor Zahnlauten unverändert, vor anderen Lauten in den ihnen entsprechenden umgewandelt, zum Anfangsbuchstaben des Verbums. Denn in allen genannten Sprachen gehört, wo diese Veränderung vorgeht, der Endnasenlaut des Präf, zur ersten Sylbe des Wortes, und trennt sich von der des Präfixes" (1838: 98-99). Go back to the text.

10. After expounding his view that the consonant alternation was connected with and arose from prefixation, Humboldt then stated that he believed "dass dies aber jetzt nicht mehr im Bewusstsein des heutigen Volkes liegt, sondem dass dieses in dem veränderten Anfangsbuchstaben gar keinen Zusammenhang mit irgend einem Präfix ahndet" (1838: 92). Once again it is clear that Humboldt considered the consonantal alternation with no overt prefixation as a fundamental characteristic of Javanese in contradistinction to other related languages. Thus in his brief characterization of Javanese he says: "Im Formenbau und in der grammatischen Fügung hat die Sprache ihre bestimmteste Eigenthümlichkeit in der Vernachlässigung der Verbalpräfixe und der durch ihre Abwerfung entstehenden Veränderung der Anfangsconsonanten" (1838: 321). Go back to the text.

    REFERENCES

    Arens, Hans. 1969. Sprachwissenschaft: der Gang ihrer Entwicklung von der Antike bis zur Gegenwart. Zweite, durchgesehene und stark erweiterte Auflage, Orbis academicus, 1/6. Freiburg & München: Karl Alber.

    Chomsky, Noam. 1964. "The logical basis of linguistic theory." In Lunt 1964: 914-978.

    Chomsky, Noam. 1966. Cartesian Linguistics: A Chapter in the History of Rationalist Thought. New York & London: Harper & Row.

    Gericke, Johann Friedrich Carl. 1831. Eerste gronden der Javaansche taal, benevens Javaansch leer- en leesboek, met eene woordenlijst ten gebruike bij hetzelve. Batavia: Lands-Drukkerij.

    Humboldt, Wilhelm von. 1836. Ueber die Kawi-Sprache auf der Insel Java nebst einer Einleitung über die Verschiedenheit des menschlichen Sprachbaues und ihren Einfluss auf die geistige Entwickelung des Menschengeschlechts. Erster Band. Berlin: Druckerei der Königlichen Akademie der Wissenschaften. Reprinted in 1960 as Dümmlerbuch 8356 by Ferdinand Dümmlers Verlag, Bonn.

    Humboldt, Wilhelm von. 1838. Ueber die Kawi-Sprache auf der Insel Java. Zweiter Band: Fortsetzung der Kawi-Sprache, Malayischer Sprachstamm im Allgemeinen und dessen westlicher Zweig. Berlin: Druckerei der Königlichen Akademie der Wissenschaften.

    Kiliaan, H. N. 1919. Javaansche spraakkunst. 's-Gravenhage: Martinus Nijhoff.

    Leitzmann, Albert. 1903-. Wilhelm von Humboldts gesammelte Schriften, edited by Albert Leitzmann and the Deutsche Akademie der Wissenschaften zu Berlin. vol. 7, part 1. Berlin: B. Behr.

    Lunt, Horace G. 1964. Proceedings of the Ninth International Congress of Linguists, Cambridge, Mass., August 27-31, 1962. Janua Linguarum, Series Maior, 12. The Hague: Mouton.

    Teselkin, A. S. 1963. Drevnejavanskij jazyk (kavi), Jazyki zarubezhnogo Vostoka i Afriki. Moskva: Izdateljstvo vostochnoj literatury.

    Zimmer, Karl E. 1968. "Review of Cartesian Linguistics, by Noam Chomsky." International Journal of American Linguistics 34: 290-303.