Adversative Conjunctions¹

by Otto Jespersen, 1933

The etymologist will generally consider his task fulfilled and his mission accomplished, once he has succeeded in finding a word in an ancient language which, from the point of view of phonetics and signification, agrees with the word he is desirous of explaining, so that he can set forth the reasons for the changes it has undergone in the course of time or, at least, point out similar instances in the same or other languages. The question why an older word should have been ousted by a newer is less frequently raised and yet it often involves problems which it is worth while probing into more deeply.

This article will deal with a series of adversative conjunctions on whose origin sufficient light would seem to have been thrown - enough at any rate to satisfy the etymological dictionaries in general use. In a recent short article on 'Le renouvellement des conjunctions', (Annuaire de l'École pratique des Hautes Études 1915-16), Meillet has discussed several of the usual causes that lead to conjunctions being replaced in the course of time, in that they are no longer felt to be forcible or bulky enough: he has not, however, specifically grouped together the words which will be commented on in this article and has therefore not discovered the ultimate reason why just those words and those forms have ousted the older.

We have first the familiar fact that, in the Romanic languages, the Latin sed is replaced by magis, Ital. ma, Sp. mas, Fr. mais. The change in meaning causes no difficulty; from 'more' it is no great distance to 'sooner' which, like Germ. vielmehr and Eng. rather, is well adapted for use in statements implying correction or contrast. (As regards this change see Tobler, Vermischte beiträge, 3. 2. ausg. p. 78 and Richter, Zeitschr. f. rom. phil. 39, 656).

Next we have the Scandinavian men which came into existence in the 15th century. The explanation usually given, that this word arose through a combination of meden (now medens), in the shortened form men, (cf. mens), and Low Germ. men (= Fris. men), 'aber, sondern', seems unexceptionable.

In early Middle English, at the time when OE ac was still in use, (ac, ah, auh), and but (OE butan), had not yet come to be extensively used, we find a word me which occurs in a few texts only, Ancrene Riwle and some of the 'Katherine group'. According to NED this word is 'of obscure origin' and its connexion with Scand., MDu. and MLG men is regarded as doubtful. It is described there as a particle, (exclamatory or adversative), employed to introduce a question or less commonly a statement - 'lo, now, why'. In Mätzner's Wörterbuch, where full quotations are given, it is translated 'aber' and this translation seems to fit in all the contexts in which the word occurs.

As far as the MLG word is concerned it is explained by means of niwan meaning 'only, merely', arising from the negative ni and wan 'lacking' (cf. ON vanr). 'Die bedeutung "aber" hat sich aus "ausgenommen" entwickelt' (Falk & Torp). The assimilation nw>m is easy to account for from the phonetic point of view (on this see among others Schröder, Indogerm. Forschungen 22.195 and 24.25).

We meet with the same sound shifts in the third word, Dutch maar, Old Fris. mâr from en wâre, 'it could not be', the same combination which has become nur in German.

Thus we have three distinct ways of arriving at the adversative meaning, in none of which those familiar with the changes of meaning of words in different languages will find anything unusual. But why should new words have been resorted to? Were the old not good enough?

Here I shall point out two features common to all these conjunctions. The first is syntactical: all three are placed at the beginning of a sentence; in this they differ from synonyms such as the Latin autem, or Germ. aber, which can follow one or more words. The second is phonetical: magis, men and maar, in contradistinction to the words they oust, begin with m. We find the same two features outside the Aryan lamguages in words with the same meaning, 'but', concerning whose origin I can say nothing, Finnish mutta(2), Santal menkhan. (Heuman, Gramm. studie över Santal-språket 69). We may compare also Kutenai ma 'but' (Boas, Kutenai Tales, Bureau of American Ethnol. 59, 1918, p.94), mi'ksa'n 'but' (ib. p. 98).

The explanation is undoubtedly as follows. The sound [m] is produced by keeping the lips pressed tightly together, while the tongue lies quietly in the lower half of the mouth and the soft palate is lowered so that the air can escape freely through the nostrils. Now this is the characteristic position taken up by the organs of speech of a man who is deliberating a matter without saying anything - the only difference being that his vocal chords also remain quiet while in the enunciation of [m] they are set vibrating.

How often it happens that one wants to say something, even knows that one must and will, but it is not quite clear as to what one is going to say. At this moment of uncertainty, when the thought is being born but is not yet clothed in words, one nevertheless begins the activity of speech: the vocal chords are set vibrating, while the lungs expel the air and, as the upper organs are precisely in the position described, the result is [m]. This is written hm!, even if there is no [h] (or more correctly voiceless or breathed [m]), before the voiced [m], and this is just the formless interjection of protest (3). Not infrequently an [m] of this kind comes immediately before a real word so that we get for example, Dan. mja, Eng. myes as a hesitant objection. (See Anker Larsen, Vises st. 411 M-jo, jo det kan man godt sige | ib. 149 M-nej | Hamsun, Segelfoss 38 M-nej | Garborg, Mannf. 91 Mark Oliv stauk upp haaret og var alvorleg; mja -, sa han | Shaw, Misalliance. 154: M'yes | Galsworthy, Swan Song 183 M'yes | Rose Macaulay, Potterism 50 M'm yes | Galsworthy, Maid-in-W, 43 M-no.

Now we can understand why words beginning with m are so frequently chosen as adversative conjunctions. The starting point is this sound and then recourse is had to some word or other that has some sort of meaning and which just happens to begin with the same sound: ma, mais, maar. The Danish men may well be an [m] that has slid into the old conjunction en in the same way as mja is m+ja. This is in no way incompatible with the notion that some of the first to use the word had in mind men = meden, while others thought of Low German men. We have thus three etymologies for this word which, far from being mutually exclusive, have co-operated in making it popular and common. (Incidentally, may it not be possible in the case of other words for which etymologists suggest various explanations so that one writer challenges the interpretation put forward by the other, that a similar point of view holds good so that both interpretations are correct, in that the word arose in one way amongst one group of speakers and in another amongst another?)

In the combinations mais oui, mais non, so frequently met with in French, is it not possible that, in some instances at least (4), it is only this deliberative [m] that is being uttered, before the speaker is certain whether he means 'yes' or 'no'? Mais oui is very frequently pronounced [mwi] and not [mEwi]. In the same way remarks are often introduced by [ma~fE~], written in books as mais enfin, though this expression does not always convey the full force of objection that, strictly speaking, is implied in the word mais.

The ma meaning 'but' found in modern Greek and Serbian is explained as a borrowing from Italian, in spite of the fact that contact between the Italians and their eastern neighbours has not otherwise been close enough to lead to the adoption of form words from Italian by these peoples. But precisely this word has found its way into these languages because it began with the universal deliberative [m] - or is it possible that it came into being spontaneously in these languages? The Roumanian of Transylvania also has a ma 'but', according to a communication by Sandfield. The Nigger English of Surinam shows a ma with the same meaning (in the British Bible Society's translation, and in Pikin spelle en leri-boekoe vo da evangelische broeder-gemeente, Paramaibo, 1849). Since, however, H. R. Wullschlägel in Deustch-negerenglisches wöterbuch (Löbau 1856) and H. C. Focke in Neger-English Woordenboek (Leiden 1885) have both the forms ma and mara, it is most likely that they are derived from the Dutch maar. In Die nywe Testament ka set over in die Creols taal, (Copenhagen 1818), the form maer is used.

An attempt at another word for 'but' beginning with m is found in the Greek mâllon, which Bréal has noted once only in a tabula devotionis (Mém. Soc. Linguist. 7. 187). The change of meaning is precisely the same as in the case of the Romanic magis.

If in English the word but meaning originally 'without' has taken the place of OE ac, the change of meaning is the same as that in the Swedish utan. Yet I am inclined to think that the closing of the lips at the beginning of the word has been a contributory factor, (mbut is not unknown), even if the soft palate is raised to pronounce b, so that in this instance the point of departure is not the position of complete rest. Something similar applies to Spanish pero, Ital. però (from Latin per hoc). The French bah is frequently used in a manner suggestive of repudiation, and similar interjections are met with in many languages. Berneker, Slav. etym. wb. p. 36, is scarcely justified in differentiating between the ba ('ja, freilich, allerdings'), that occurs in several Slavic languages and which he associates with avest. ba, 'partikel der beteuerung un hervorhebung', etc., and Russ., Bulg., Serb. ba, 'ausruf des stauens', which is taken to be a 'primäre interjektion wie nhd. ba, frz. bah, osm. ba'. These words are identical and serve as more decisive subordinate forms of the rather more uncertain and hesitant ma.

There are other words beginning with m which one is tempted to mention in this connexion since they express something similar to the reflective, half or wholly reluctant [m]; especially verbs like E. mope, Dan. måbe. Further E. mumble, mump and mutter, Dan. mumle, Lat. murmurare, as a means of expressing partly weakly spoken sounds and partly the unformed objection. We have many similar words with related meaning: Dan. murre, mukke, Germ. mucken, mucksen, muckern, adj. muksch, muckig, 'peevish, grumbling': in Dan. we have, with the same meaning, the adj. mut with its subordinate form but; otherwise m- and b- are not interchangeable in this way: we can understand their being so in the case of these words because here the most important thing is to have the lips closed at the beginning of the words. English has the word moody with the same signification; here etymologists no doubt refer it to OE modig and explain that mod means 'state of mind' generally, not only 'courage' as Dan. mod, but also a distinctly unwilling mood. Yet that moody should have come to mean 'sulky, sullen' and thus acquired the same unfavourable shade of meaning as Dan. mut is undoubtedly connected with the use of [m] commented on in this article. We may also compare Eng. mum, 'silence, silent, quiet' (originally, not as is stated in NED an 'inarticulate sound made with closed lips', but rather, the sound emerging when one begins to speak first with closed lips, then opens them and at once breaks off speech and closes them again), with the remarkable mumchance which is used like the Dan. mut, (Locke, Ordeyne 174 I sat mumchance and depressed).(5)

Dan. mukke leads us to muk used in the negative ikke et muk 'not the slightest sound', Germ. keinen mucks and thence again to Fr. mot, Ital. motto, Gr. múthos, all of which have acquired a more exalted and complete meaning than that contained in the Danish and German half-words.

The reason that Dan. mon (almost always the first word in a sentence) from being a form of the verb munu, has come to be an interrogative particle is undoubtedly the fact that, with its initial m, the word was well adapted to begin a dubious, hesitant question: mon han kommer? (originally ... komme) is rather more uncertain than kommer han?, and is this quite naturally introduced by a hesitant [m]. We may compare the Gr. môn with the same meaning, used in introducing a question, even if, as seems probable, the Greek word is derived from mè oûn and is thus quite different in origin from the Danish: the similarity between the two cannot be ignored, it is of the same kind as that between Ital. ma and Dutch maar.

While in the instances already cited we have an initial closing of the lips, in American nope [no·up], [no·p] and yep [jep], we find a final closing of the lips of somewhat similar value, usually without audible explosion, but I do not clearly understand what feeling prompts the use of these variatons of 'no' and 'yes', although I have often heard them. In literature they are found not only in modern American, but also in recent British authors.

In the beginning of an utterance we have not only [m], but also [n], since in moments of silence the tongue often rests in the advanced closed position while the nasal passage stands open; whether at the same time the lips are closed or open does not matter at the moment that the vocal chords are set in vibration, the audible result is in any case [n]. This is used in the same way as [m] before yes: E nyes, Dan. nja. cf. the author's Fonetik p. 272, Lehrbuch der Phon. 5.62. Now it is worth noting how many words meaning the same thing as men there are which begin with n-: Russ. no and specially French néanmoins, Ital. nondimeno, Sp. no obstante, Germ. nichtsdestoweniger, long awkward words whose length is just adapted to give time for the coming objection to take shape, since, when one wishes to contradict the person with whom one is talking, it is important not to hurry, but to weigh one's words lest they give offence!

The last mentioned words contain the negation as the first element and since negative words begin with m and n not only in our own family of languages (ne, me, etc.), but also in many others, Magyar, Eskimo, Sumerian, Duala, Arab, Egyptian, Chinese, we are surely justified in seeing in this an allied outcome of the same tendency to want to say something with the organs of speech in a resting position. Only in the case of the words just cited the word issuing is to a still greater extent a querulous, repudiating one, originally probably an expression of refusal, disgust, aversion, which is also conveyed by means of what we describe as 'turning up one's nose'.

In the last sentence I have said nothing new, but I do not think that the words for 'but' alluded to in the beginning of this article have been accounted for by others in the way that they are here. In conclusion I will point out that my explanation is not based on sound symbolism in the ordinary sense of the term. In my opinion the words are certainly in one sense natural words but entirely different from 'onomatopoeias' or echo-words: they are the usual type of words, (conventional), but have come to be used in this special way because they contained an element deep-rooted in human nature. They may consequently be regarded as secondarily natural words. The ancient Greek philosophers debated whether words arose phusei or thesei and could only imagine the one or the other origin, we see here in one individual province a union: the words have arisen both phusei and thesei.

Linguistica, 1933.

(1) In Danish, Nogle Men-ord, in Studier tillegnade Essais Tegnér 1918. Here translated with some additions and slight alterations.

(2) Vilhelm Thomsen wrote to me about this word, (April 4, 1910) 'As far as I know mutta occurs only in Finnish and, borrowed from Finnish, in Lappish mutto. It is not even found in Estonian (aga), still less in any of the other more remote Finno-Ugrian languages. In all probability it is connected with muu 'other', although I am not quite sure how it came to be formed. The particles often present remarkable developments. Some individuals are said to show a variant I do not know, muutta, with long u, but whether this is a legacy from the original form or is due to analogy for example with muutoin 'otherwise' or some similar word I do not know.'

(3) Other ways of writing this interjection hem, hum (see NED); Um (Kaye Smith, House of Alard 295, 300); Mm (Lawrence, Ladybird 126 cf. 194). This word like a great many others is ably discussed in Hjalmar Idforss, De prim%auml;re interjektionerna i nysvenskan. I. Lund, 1928, but the author seems rather too much inclined to suppose conscious literary loans from one language to another and to underestimate the essential uniformity of human nature in all nations. What is taken over may only be the fashion of writing down sounds which have been pronounced and heard from time immemorial.

(4) In other cases mais oui, mais non, is emphatic as in jolie, mais jolie! (for this observation I am indebted to Schuchardt).

(5) Cf. H. Petersson, Verg. slav. wortstudien 50 lautegebärde mu (1) mundverschliessen, Gr. múo, (2) leises bewegen der lippen oder ein murmeln. - Cf. further E. muzzle, Fr. museau, OF musel.


Back to International Auxiliary Languages

James Chandler 03-Aug-01