The Analogist and Anomalist Controversy F. H. Colson The Classical Quarterly, Vol. 13, No. 1. (Jan., 1919), pp.

24-36.
Stable URL: http://links.jstor.org/sici?sici=0009-8388%28191901%291%3A13%3A1%3C24%3ATAAAC%3E2.0.CO%3B2-J The Classical Quarterly is currently published by The Classical Association.

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/classical.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers, and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take advantage of advances in technology. For more information regarding JSTOR, please contact support@jstor.org.

http://www.jstor.org Wed Aug 22 11:09:10 2007

T H E ANALOGIST A N D ANOMALIST CONTROVERSY.

THE controversy between the Anomalists and Analogists has not, I think, attracted as much of the attention of scholars as it deserves.' It was perhaps not a very practical matter, though, as I shall point out presently, it probably had indirectly some important practical results. The interest of the controversy lies rather in the spirit in which it was conducted. Anyone who reads for instance Varro, De Lilig. Lat. VIII. 31-32, where the anomalist argues that as in life variety of furniture and the like is necessary for aesthetic enjoyment, so in language anomaly is desirable ; or IX. 24, etc., where the analogist argues from the unchanging order that prevails in the heavenly bodies, in the tides, in the continuity of species, will feel that he is moving in a world of thought very different in one way from our own, though in another rather like it. By the analogist language is conceived as a world in itself, much as we conceive of the visible world. Its phenomena are being laid bare and constantly reveal fresh signs of law and order. The investigator sometimes finds facts which prima facie suggest anomaly, but he is as confident that behind them must lie some unifying principle a s the scientific man of to-day is with regard to the phenomena of the visible world, as impatient of the suggestion of disorder as he is of any miraculous interference with the order of nature. Even, the anomalist, sceptic as he is, approaches the question not in a spirit of mere denial, but of aesthetic consideration. We get a glimpse of a lost point of view. The world of words had a glamour and a wonder for them which it cannot have for us. When Greek philosophy first turned to the subject of language, the earliest distinction it set up was between ~b ~ ~ ~ ~ ~ and v 6 rb ~ u7jpurvov, v o v the thing meant and the actual word spoken. When it made the former the primary object of study it produced the theory of the parts of speech, and its final results in this department, and indeed the terminology itself, have come down to us practicallyanchanged. When it made the latter its primaty object it produced 'etymology,' of which unfortunately the same cannot be said. A little further consideration of words however led to the conclusion that their relationship varied &eatly. Some words were obviously very closely related, so much so that they could only be regarded as flexions or KUUSLF of each other. In other cases, though a relationship might be suspected, it could
1 Much of this paper covers the same ground as parts of Ste~nthal's G~schrchtc &r Sprachrmsonshalt bn dcn Enrchcn und Romcrn. The greater

part both of the facts and of the inferences has, I think, been arrived at independently:

T H E A N A L O G I S T AND ANOMALIST C O N T R O V E R S Y

25

not be affirmed with equal certainty, while in others no relationship was discernible. T h u s a distinction arises between the word itself (which in nouns was naturally identified with the nominative singular; in verbs, with less reason, with the first person singular present), and its KXIUELF. I t is t o the science of the former that the term E ) r u P o h o y ~ a becomes restricted. T h e science of the latter never really acquires a name, but has to content itself with employing a term which arose in a diiferent quarter-analogy. U'l?ilst philosophical thought was concerning itself ~ v i t hthe nature of language, the grammarians were busying themselves with the correction of texts, and particularly Homer, who, it must always be remembered, was to them to some extent a writer in a dead language. In determining doubtful points, the principle of 'proportion' or ' analogy' was one of the leading tests. If the genitive plural of B&F was according to some Bdwv, according to others Ow;,, Ivhile the genitive of Bdp was undoubtedly B.rlpoiv, it was reasonable t o i :vB&F : B?jp.l I see no reason to doubt that this argue that BwGv : B ~ ~ p : process u.as applied from very early days o i grammatical investigation, or (in spite o l Steinthal) that Aristophanes himself attempted to formulate the condltior~s under which analogy could be guaranteed. Analogy then is properly the name for the principle which was used in the science of KXIULF rather than for the science itselfa2 Rut it comes to be used for the science itself. T h u s in Dionysius Thrax and Varro, we find it regularly as the counterpart of E)rv,uohoyla. Both of these sciences had their opponents as well as their defenders. Varro devoted three books (now lost) t o 'quae contra etymologiam dicercntur, quae pro ea, quae de ea,' in just the same way as he treated analogy. But naturally the controversy on analogy, which claimed an order and method, which etymology could not claim, was far warmer. I n this controversy, I imagine, the tendency of the grammarians as a whole was to analogism. T h e scientific spirit, as I have said, was naturally impatient of disorder. T h e opposition seems to have come, to some extent at least, from Stoicism. T h e leading anomalist, Crates, at any rate was a Stoic, and he founded his views, even if by misapprehension3 as Varro says, on Chrysippus' treatise m p i & v w , u a X l a ~ . I t is at first sight puzzling to find the Stoics on the side of anomaly. W e know that they held that language was due^ and that words corresponded t o their meaning, and we hardly expect to find them on the side of disorder in K X ~ O L F . Perhaps however the reason is not far to seek. If we examine the passage in (pseudo- ?) Augustine, De dialectic^,^ which seems t o be our chief authority on the subject, it does not appear that the Stoics held In the first place, logical that words, as they actually existed, were r#due~. or illogical processes have deflected words from their natural meaning : ' Vis '
So Aristarchus, quoted by Herodian, on
11. XIII. 103 ; u. Lehr's Ar~starchus,p. 261. Cf. Pindarion apud Sext. E m p adu Gramm. 10. zoz : draAoyio. iurlv ipdov re ral dvopoiav
3

'

' This passage is perhaps most accessible in

dc L. L. IX. I.

...

Brwpio.

Goetz and Wilmann's edd. of Varro ; whether there are adequate grounds for the implied view that the writer draws from Varro I do not know.

26

F. H. COLSON

is a natural word, for the ietter ' u ' has a forcible sound. Thence we get

' uinculum,' because chains use force (per efficie~ztiai~z), and hence ' uitis ' and
' u i a ' ( p ~ similitudinciit), r because vines and roads are apt to wind. At this point Augi~stinehas had enough of it, and concludes ' quot modis autem origo uerborum corruptione uocum uarietur, ineptum est persequi.' I understand this covrz~>tio zroctriiz to refer t o the other questions which the Stoic must have had to answer. Granted that ' uis ' is + u o e ~ ,that ' uinculum ' followspev e-ficientializ and ' uitis ' per si~~zilitztcline~i~, it must still be asked, whence comes the rest of the word ' uinculum ' ? T h e Stoic presumably fell back upon the ' demptio, additio, traiectio, commutatio' of letters and syllables of which Varro speaks.' In fact, though words may be to the Stoic of nature, it is nature corrupted. And this view is, I think, supported by the word 'etymology,' a rather curious term when one comes to think of it. I t does not of course mean ' t h e truth about words,' but ' the science of h u p a . ' T h e doctrine that behind words, as we have them, lie roots which are true things, implies that as we find them they arc not true. If this is the Stoic view, it is not surprising that Chrysippus and Crates, having found disorder in language up to the stage of K X I U L F ,should not be inclined t o recognize immutable order in x h i a ~ sitself. T h e strength of the anomalist attack fell most powerfully on what we may call the outworks of KXLULF. T h e earliest conception of x h i o ~ cdoes not distinguish between inflexions and what we (perhaps unconsciously following a line of thought laid down by analogistic grammarians) should call derivatives. In Varro ' R ~ m a is '~ a declinatum of 'Romulus,' 'ouile' of 'ouis,' while 'equitatum a b equitibus, equites a b equite, equitem a b equo,Ia is spoken of as a homogeneous series. T h e anomalist laid stress on the great variation we find in such formations. W e say ' o ~ i l e , 'but ~ not ' bouile '; ' uinaria,' but not ' carnaria ' ; ' Romanus,' but ' Parmensis '; ' cantito,' but not ' amito '; ' aucupem,' but not ' piscicupem.' I t is obvious that all this sphere of language does not lend itself t o systematizing either in Greek or Latin, and the analogist's answer was to surrender the whole province. ' No one guarantees that a word in the nominative singular should be declined into another word in the nominative singular on the principle of a n a l ~ g y . ' ~ Varro himself rests this on a general principle. W e must distinguish between declinatio uoluntaria and declinatio naturalis. T h e former rests upon the caprice of men and exhibits inconsistency ; to it belong such formations as ' Roma ' from ' Romulus,' and ' Tiburs' from ' Tibur.' On the other hand, ' Romulus '-' Romulum '' Romuli ' is decli~zatio natztralis. Varro does not definitely say that ' ouile,' ' uinaria,' and the like are declinatio~tesuoluntariae, but I think it is implied. W i t h this element eliminated, the field was restricted t o KXLULF proper. Here one of the main points of attack was defectiveness. T h e anomalist pointed out that some nouns had no plural and others no singular, that 'bonus' had no comparative and ' melior' no positive. Again, they drew
Ds Ling. Lat. V. 6 .
2

Ib. VIII. 18. Ib. VII. 4.

I b . VIII. 54-62. Ib. IX.50.

T H E A N A L O G I S T AND ANOMALIST C O N T R O V E R S Y

27

attention to ' Iupiter '-' Iouis,' and the like. T h e answer of Varro's analogist is that the first class of defectiveness arises either from nature or from use. I f a plural cannot be conceived of as in 'gold,' the word cannot exist ; l if a n animal plays so little part in human life that its sex does not concern us, we cannot expect both masculine and feminine forms. W e used to call all pigeons ' columbae,' says Varro or his analogist authority, but now that the keeping of pigeons has become common, 'columbus ' has come into use also.2 W i t h regard to ' Iupiter '-' Iouis,' etc., the answer given is that such cases resemble the statue which has lost its head, or had an alien head fitted t o it. Put Philip's head on Alexander's statue and the proportions of the rest of the statue are not a f f e ~ t e d . ~ With regard t o a somewhat different point, Latin masculine nouns in -a', the analogist declared that this difference of meaning did not affect the analogy. They were really still feminine nouns. h man may sometimes wear a woman's shoe, but it remains a woman's shoe for all thate4 But the main controversy gathered round nouns of similar nom~natires but different declension. T h e opponents of Aristarchus had cited Qrhop48rlv and 'HpanXcIS1/~ ; Aristarchus answered that the vocative must be taken Into account as well as the nominative before the words could be postulated as similar, and therefore t o be declined analogically. T h e anomalists replied that a contention of this kind made nonsense of the whole controversy. YOU I4y down that, given similar parents, the offspring is similar. You find similar parents, observe the offspring, and find them dissimilar, and then infer that the parents are really d i ~ s i r n i l a r . ~ T h e analogist replies that it is rather t h e case of two appearing alike in a poor light. You bring in a candle and find that they are really unlike. H e illustrates it by the nouns in - x , as 'crux ' and ' Phryx.' They appear to have the same final letter, but ' crucis,' ' Phrygis ' show the real d i f f e r e n ~ e . ~O r again it is like the case of Gallic and Apulian wool. T o the unexpert they seem the same, but the expert knows that the Apulian is the stronger, and pays a higher price for it, and so too the real similarity of words must be tested by their effect.' T h e same may be said of verbs : ' dolo ' and ' colo ' seem the same, but go on to ' dolas ' and ' colis ' and the difference becomes clear." This principle of A r i s t a r c h ~ s which ,~ regards the noun a s a whole, and does not start from any particular case, is of course the only sound basis for a theory of analogy. But it did not satisfy the grammarians as a whole. They were too closely wedded to the idea that the nominative of the noun was its ' prima positio,' the real noun from which the cases were merely ' fallings
L.L. IX. 66 a Ib. IX. 56. Zb. IX. 79. Ib. IX. 40. I b . VIII. 68, 69. 6 Ib. I X 43,44. 7 Ib. I X . 39. 8 Zb. IX. 108. T h e controversy seems t o have dealt with the noun a great deal more than with the verb. T h e analogists seem to have argued that formations from the present stem belonged to a different word than formations from t h e perfect stem. If 'lego,' ' legam,' ' legebam,'
1

3
6

'

follow analogically, and also ' legi,' ' legero,' 'legeram,' we have no right to expect analogy between ' lego' and ' legi.' If this were once granted, the regularity of the verb in Latin a t any rate is very constant, and I suppose offered less target for the anomalists. 9 But perhaps Aristarchus h~mselfwas under the spell of the nominative, and chose the vocative as being merely a variat~on of the nominative.

28

F. H . COLSON

away.' They accordingly rather took the line of showing that where there were differences of inflection there were also real differences in the nominative. The simple and general rule for the comparison of any two words was, a s Varro gives it,' that unless the ' casus, genus, species, exitus,' were the same, they could not be expected t o correspond. I n this forrnula cnsz~sapparently ss of course gender, species means part of speech includes number, g e i ~ z ~ i (adjectives not being distinguished from substantives, but on the other hand or t t rrpouvthe ?loillen or Buopa n i p r o u being distinguishrd from the z i o c n b : ~ l z ~ ~ y o p i a ) . E x i t z i s is of course ' ending,' but it was clear that if this was restricted t o the ' litterae quae commouentur,' the rule was useless to the analogist in his controversy with the anomalist, who could point out hundreds of instances where these conditions were satisfied and yet analogy did not result. Varro is therefore obliged to add that not only the 'litterae quae commouentur,' but the adjoining letters must be taken into consideration, for ' haec u i c i n i t a s aliquantum p o t e ~ t . ' ~ This vague suggestion had naturally to be defined more closely. Some grammarians (perhaps only Latin) laid down the rule that the letter before the exitus, and therefore the whole of the last syllable, must be the same. This too was probably found inadequate. A rule quoted by Charisius, as originated by Aristophanes and supplemented by Aristarchus, adds to the above-(I) number of syllables, (2) accent, (3) that the words compared must be both compound or both ~ i m p l e . ~ But by far the fullest and clearest rule is found in the following passage, which is attributed to the grammarian 4~ c i p l e p y i j Hcrodian : Ti, i ; p o ~ o u i u r o t 7 d u 6 p a u ~ u4 yElue~ 4 E%EL jj o x 7 j p a ~
76uy rrr&uer,

4 xaraX7jfe~

QU

vaparcA~6~ yX X a f i $ uv

I u x p d u y , du

T O U ~ T ~ T L

Ij yiyu~ra~T . h e grammarian goes on t o take various pairs of words of different declension and to show that they fail in one or other of these ~ +LX~T~ are P different in gender. ' O ~ U ~ X L Oand YIK~F respects. T o f d ~ qand I I o X u u I x q ~are of different elfiov, for one is a r r p o u q y o p ~ a ,t he other an b'uopa ~ ~ C o x p d r q r are different in u x i j p a . One is simple, the and ~ 6 p ~ o v '. I r r ~ 6 . r other compound. ' I x B G r i ~ B 6 o r and lXBDr i x 8 6 o v are different numbers. " H p o r and e 8 p i 7 have different accents. Tot67177 and Q X C ; T ~ F a re different cases. K a X 6 r and fipa86r have a different ending. I I 6 p u q ~and A d x q v are of different length in the penultimate, ' A p x n ' r and lp& in the final vowel. A v o i a r and /3Iav have a different number of syllables. Z w X . j u and ;,piu have a different consonant in the last syllable, for p often turns q t o e.6 This last case is put
~ ~ U ~ ~ O L KXIUIF O F
1

L.L.X. 21.

a Ib.

X. 26.

3 Gram. L a t . ( K ) I . 127. I do not think the vagueness of Varro's rule, as compared with that of Charis~us,i s conclusive that the latter is not a s old as Aristophanes. Varro may well have seen that accent a t any rate cid not fit Latin, and hesitated at putting out a definite Latin rule. I print the extract as given in Cram, Anec.

'

O x . IV. 333. 6. But the colon after vvAAaj3ij is clearly a mistake, as also the absence of comma after xaiaAd[r~. T h e rardA?Er is the exctus. T h e sapeAiyovra ovAAaj3ij is .the penultimate. T h e series of 4 gives the main considerations, the series of i v subordinate ones. 6 T h e phrase irrrrAori, rup)$vou means, I suppose, the tendency of t h e consonant to influence the following vowel.

T H E ANALOGIST AND ANOMALIST C O N T R O V E R S Y

29

forward somewhat tentatively, as a thing which does not happen regularly but often. If you cannot find any other difference in two nouns which are differently declined, you may fall back upon these consonants as the last resort. T h e anomalists probably did not accept this law of uicifiitas, whether in its later or its earlier form. But no doubt they preferred to meet the analogists on their own ground and beat them there if possible. Their trump card (in Latin) was ' lepus ' and ' lupus.' ' Lepus ' is of course an anomaly, being the only masculine noun in -us belonging to the third declension, but the anomalists strengthened their case by pitting it against 'lupus.' Here they said every possible demand of $&initas is satisfied. T h e two words are identical except in their first vowels, and even these vowels are both short. T h e analogist was fain to reply that the two really differed in gender, for 'lepus' is epicene, having no feminine.' So again the anomalist put forward ' aper ' and ' pater.' Here too it was undeniable that there was no difference of uicinitas. The other side replied by extending the idea of species. Grammarians in classifying nouns had distinguished between 'relative' and 'absolute' or 'positive' nouns. Since a father implies a child, ' pater ' is a relative noun. Not so ' aper,' and hence the difference of declension does not contradict a n a l ~ g y . ~ These analogist apologetics are of course very futile and even pitiful. If we take the tests of the alleged Herodian, gender, number, case and ending are obviously right. Perhaps too accent and even uxfipa as bearing on accent have some sense. But the other tests, such as number of syllables or the length of the penultimate syllable, have no relation to reality so far as I can see. They served to baffle the anomalist because, whenever he pointed out a difference in the declension of two words of similar termination, it was generally possible to find one of these accidental differences. But at the same time they really obscured the true laws of analogy. When the grammarian says EL 62 ~apT & UCTL T L y & v ~~ ~ V ~ ~~ ~ O , L 6OK F~ ~ U Ly Fh y ~ ~he ~a destroys ~, his whole case, and indeed talks such obvious nonsense that I can hardly believe that he realiy represents the views of Herodian. Still on the whole it seems to me that the analogists were right. They were right, that is, in maintaining that order reigns in KXIULF.W e know as a matter of fact that a knowledge of comparatively few rules and examples will enable the student of Greek and Latin to decline correctly a very large number of words. \Yhen the analogist justified his position by the order of the heavenly bodies he went too far. But the passage at the end of Book IX., where Varro's analogist remarks that declensions differ from each other as the species of animals differ, but that the similarities within the classes are far greater and more striking than the dissimil'arities of the classes, and that the
Qulnt. I. 6, 12. Qoint. 1. 6, 13 This example brings us a good way on to the idea. that analogy depended on the meaning of w0ri.s. I n general the analogist repudiated thls doctrine (L.L. IX. 40), but we find traces of some such notion. Probably a t any rate anomalists felt their case to be
1 1

stronger if they could allege similarity of meaning. T h e iorce of lefus ) ( l~rfi~i was, I suspect, increased by the fact that they are both wild animals. In L.L. VIII. 10 we have associated with ' lepus' and ' l u p u s ' the odd word ' surus.' This has been amended to ' s(ci)urus.' but I should suggest 'ursus.'

30

F. H. COLSON

occasional exception stands on a par with the freak or deformity, the 'equus claudicans ' or ' homo luscus,' he makes a just comparison. As he says, you may apply t h e term 'anomaly ' to the variations in nature, yet no one denies that order reigns in the animal and vegetable world, and so it is in the realm of KX;ULF. There is another point of view which ought to be noticed. This is t h e view that analogy itself, so far from being the opponent of 'usage,' itself springs from usage. I t appears in the opening chapters of Book VIII., in a part of the work where Varro certainly seems to be giving his own opinion, though it is not easy to reconcile it with his oft-repeated statement that , ~ says analogy is of nature.= But it is more fully developed by Q ~ i n t i l i a nwho that analogy was not something heaven-sent which dictated the form of speech, but is what we find that men actually do speak. I t therefore rests on exeii2)lun1, not on m t i o , and is actually created by usage. This view is in the main, I suppose, anomalistic. I t abandons the idea that analogy must or even ought t o control consuetzrdo. I t is not, as Quintilian says, a lex which must be obeyed, but an obsertratio. Still it is so far analogistic that it does not deny that analogy is the dominant factor in KXIULF, and where covsuefzuio is doubtful it admits the presumption that analogical forms are t o be preferred. F o r practical purposes-so far as practical purposes were involved a t all-those who held it might well be analogists in tendency. This brings us to the question, what were t h e practical results of the controversy ? I t has been remarked by others that one result was the formation of the canons or rules of declension. At any rate the controversy was a contributory cause to the work. \Vhen we try to realize the position in which a grammarian of the second or third century B.C. found himself, we may doubt whether such an herculean task a s the reduction of KXLULF t o a system would ever have been carried out without some such initial motive. I t requires perhaps more philosophy to distinguish the parts of speech, but it requires little labour. But the analogist t o maintain his position had t o survey the whole of the language, otherwise, when he had met the anomalist successfully on one point, he might always be confronted with some overlooked exception. T o survey the whole body of the nouns and verbs of the Greek and Latin languages, with no dictionaries or grammars, even of other languages, to serve a s a guide or basis, seems t o me a gigantic business. Yet, I take it, it was done in such a way that, until the birth of comparative philology, research found little to add t o it. On the other hand, it does not seem likely that analogical theory was
1 ' Sed ii qui in loquendo partim sequi iubent nos ccnsuetudinem partim rationem, non tam discrepant, quod consuetudo et analogia coniunctiores sunt inter se quam iei credunt, quod est nata ex qcadam consuetudine analogia' The words ' ii-credunt ' are taken by Steinthal and Sir John Sandys to indicate a third party, of

which they justly say that 'till Parhm is defined, it carries us no farther.' I take the words rather, *these disputants (anomalists and analogists), who some of them bid us follow reason, while the others bid us follow usage, are not so far apart as they think.' a I. 6. 16.

T H E i i N A L O G I S T AND ANOhlALIST C O N T R O V E R S Y

31

productive of much change in the spoken language. As the analogists stoutly mairitained that language as it stood was analogical, as they attempted t o justify most apparent exceptions and admitted nothing in the end but the ' lame horse' and the ' one-eyed man,' there could be no widespread return to analogy, simply because there had been so little departure from it. Yb'ith regard to the admitted exceptions, we find in certain chapters of Book IX, chapters1 in which Varro is clearly stating the opinion of other analogists rather than his own-the suggestion that where analogy has been lost it ought to be restored. I t is admitted that this will result in novelty, but it is argued at some length, from the examples of artists, generals, and others, that novelty has often t o be raced in behalf of what is right. How far there were ~ c t n a l l y grammarians who carried out this principle on any considerable scale, who spoke themselves and tried t o make their pupils speak on strict anaiogical principles, is a point on which I should be glad of enlightenn~ent.~ Sextus certainly writes a s if there were people of this kind, who said ZQvvZ ~ O F - Z ~a Lnd K ; W U - K ~ W V O ~ - K ; W V L . ~ So too Quintilian quotes a s a happy remark about the analogists ' aliud esse Latine loqui, aliud grammatice.'4 But both these are rather vague, and such remarks need not be more than anomalist taunts. I t was a fair weapon against' the arialogists to assume that they would carry out their principles, and to point out the absurdities they w7ould land themselves in. Faillng better evidence, I should suppose that the accepted view is given by Varro himself, though he puts it in a curious way. T h e people, he says, ought t o follow analogy, but individuals are on a different footing, for they are in the power of the people. T h e poet indeed may defy custom with impunity, but the orator and other individuals cannot. So till the people come to a better mind, the orator must submit t o c o ~ t s u e t u d o . ~This melancholy acquiescence in the less desirable reminds one of the attitude of our B o o k of Com~>zolz Prayer in the Commination Service towards Open Penance. But it represents, I take it, the general attitude. Analogy is right in the abstract-even Quintilian, anomalist as he is at heart, admits this.O But ' until the said discipline may be restored again, which is much to be wished,' we must submit t o usage. T h e exception is a heretic, but the Church is not in a position to persecute.'
1 IV..IS. were differently declined, but is not recorded as T h e most definite instance I know of is the saying anything that savours of the v i e w mencase of S~senna, who used 'adsentio' for 'adsentioned by Sextus. Caesar, who wrote ' d e ana. tior' in tile senate, and was followed by others logia,' and should on the same principle be a (Gell~us, N.A. 11. 2 5 , and Quint. I. 5, 13, from 'strict adherent of analogy,' evidently argued against some analogist views (Cell. N.A. I. 10, Varro). This is not a very flagrant defiance of and XIX. 8). usage. Sir John Sandys (Htst. of Class. Schol. I., p . 142) says of Trypho ' t h a t the titles of several 3 Adu. Guam.X. 195. I. 6,27. of his books show that he was a strict adherent 5 L.L. IX. 5 , G. of analogy.' But analogy is often a mere name 8 I. 6 . 21 : ' Recta est haec uia. quis negat ? for declension (u. p. 25), and to write on it does sed adiacet et mollior et rnagis trita.' not show the views of the writer. Nor do the 7 Cf. 'Consuetude non ratione analogiae sed multorum consensione confragments, as given by Velsen, show much analouiribus par . gistic tendency. H e certainly postulated 8op and ualuit, ita tarnen ut illi artis ratio non accedat Goup a s nominatives for 8opbs-8oupbs, and resed indulgeat,' Gramm.Lat. (K.) I. 50. marked that the plural and singular of my~rXvs

'

..

32

F. H. COLSON

There was however a sphere where the analogist could move with greater freedom. This is what Pliny calls dubius sevmo, cases where usage itself oscillated more or less between the analogous and the anomalous form. Sextus tells us that the grammarians lay down the rule that simple nbuns in -79 oxytone, as E ~ U Eand P ~ eC+v+, F make their genitive in -OUT, not -ow, and that therefore the genitive of ~ ~ / . L E v $ i Ps E ~ ~ E U O O Fnot , e C / . ~ ~ u o i t . l AS there is no authority for E ~ ~ E V O the ~ , adjective, presumably he refers to the proper name Ehp~u6~, which does exhibit these variations. So too Quintilian says that the analogists supported ' domu,' ' domuum,' against ' domo,' ' domorum,' apparently on the grounds that ' domus ' as a fourth declension feminine noun found its analogy in ' anus,' ' manus.'= Steinthal may be right in saying that by the end of the first century A.D. the controversy had died down. T h e main controversy did die down, because two points were gradually established-one, that analogy prevailed in inflexion; the other that, where custom was thoroughly established, the analogist must acquiesce. But transferred to the arena of dubius sermo the contest continued, and the only thing that could bring it to an end was the decay of grammar as a living study. The questions debated and the mental attitudes displayed were the same in nature as those which we sometimes meet ourselves when such points are raised, as whether 'none are,' ' it is me,' are really ' bad grammar,' and therefore to be avoided. T h e difference lies in this-that whereas such points are to us little more than amusing trifle^,^ to the ancient mind they were questions of real and substantial interest. Whether however even here analogist theory had much influence on usage may be doubted. I t would be possible, I suppose, to see how far anomalous forms in the older language tend to give way to analogous forms. But even if it proved to be so, it would not be much evidence. Analogy, philologists tell us, is a popular instinct, which is always reshaping language, and changes in this direction are quite as likely to be genuine co~tsz~etudo as the result of conscious grammatical effort. T h e description of the opponents varies occasionally. QuintilianL couples with aiialogy etymology, as being also a ' rational ' process, in which we reason from one word to another. Etymology will sometimes justify a form which might otherwise be thought barbarous, as ' Triquedra ' for ' Triquetra,' or ' medldiem ' for ' meridiem.' O n the other hand, consuetudo is frequently associated with nzictouitas. An anomalous form may be justified on the ground that it has good literary authority, quite as much as by common usage. Quintilian distinguishes between auctovitas and uetustas. A variation of the
~

1 Ad,'. Gvamm. X. 222. I do not understand t h e word d a X G . All these words a r e surely compound, and Sextus could hardly fail to see it. I t may he objected to my assumption that he refers to the name Ebpru7js that the analogists would have said that analogy could not be expected between the uocabuilc?n r6prv7js and the nontcn Et;prv7js. But it does not follow from the e v ~ d e n c ethat either ( I ) they all accepted this

distinction, or ( 2 ) that they held it valid in t h e case of two identical words. I. 6, 4. I myself, by using the phrase ' none of your correspondents have,' became the subject of qulte a n animated controversy in a local newspaper. 4 I. 6 , I .

34

F.

H. COLSON

but only that so far as it is true it applies to the whole, and not merely to the parts. At other times it means 'universal' or 'without exception,' and is opposed to 7 ; 057 i r i 7 ; .rroXt;. I t is used in both senses by the grammarians. When Herodian's work on accents is called r s p i x a O o X r ~ ? j r .rrpoow6iar, it means I suppose ' accents as a whole,' but when, as frequently in the Latin grammarians, it means a 'rule,' it bears, a t any rate in the mouths of analogists, the other sense. Sextus brings it definitely into connexion with the analogist controversy.' T h e grammarians, he says, allege certain ~ a $ o X r ~ h B ~ w p 7 j p a r a ,by which they judge individual words. H e illustrates by the case of eiipeu7j7 mentioned above. The champion of E i i p ~ u o t , he continues, will perhaps reply by denying the ' catholicity,' while others maintain that 'catholic' merely means ~b E'K ~ X E L ~ U W U This . however is ridiculous, for x u $ o X r ~ ; u is clearly different from ~b &F E'ai 72, v o X 6 . The former never deceives us, the latter does so, though rarely. The natural inference from this is that the 'catholicon ' was in the eyes of the analogist a universal rule which admitted or ought to admit of no exceptions, while those whose tendencies were anomalistic whittled the term down to a general principle. I think we find an allusion to the first view in a passage of Quintilian (11. 13,14), where he says: 'semper mihi moris fuit quam minime me alligare ad praecepta, quae K U $ O X L K u ~ ocitant, id est ut dicamus quomodo possurnus, uniuersalia uel perpetualia.' Quintilian is not indeed speaking of grammar, but of taste and suitableness in oratory, but he does not get the term from rhetoric, nor, so far as I know, from any other art or science. I t was in grammar that he was familiar with (and rejected) laws which claimed ' universality ' and 'perpetualness.' I do not know whether it is fanciful, but it seems to me that Quintilian sees in the word, as some used it, a certain passion and insistence. Note the frequentative ' uocitant,' which I do not think he uses elsewhere, and the 'quomodo possumus.' T h e word was to him hardly translatable. I t had in fact acquired that force which acute controversy gives to a word, making any translation, even one so strong as Quintilian's doublet, seem cold in comparison with the original. T h e question naturally occurs-did this analogistic use of ~ a $ o h r ~ dhelp v to shape the meaning of a watchword second to none in its influence on the history of Christendom and Europe? The term ' Catholic Church,' as Lightfoot notes,z goes through a certain transformation of meaning in the second century. I t begins with meaning the Church as a whole, as opposed to the local churches, and the term, though Lightfoot does not note this, was almost a necessity, because in this particular case the name of the whole was the same as that of the parts. Regiments do not make a regiment, but an army. Limbs do not make a limb, but a body. I t is only E ' ~ z X q u i a rwhich make a n f ' c x X q u l a . I t came however as early as the days of Clement of Alexandria and the Muratorian Canon to acquire the sense of ' orthodox ' a s . opposed to heresy. Lightfoot explained the change by saying that the heresies were
1

Adu. Gvamm.

(K) X. 221, etc.

On Ignatius ad Smym. 8.

T H E ANALOGIST AND ANOMALIST CONTROVERSY

35

' partial, scattered, localized, isolated.' But this was not the feeling of the Church itself. Justin con~plainsthat Marcion drew his adherents from every quarter; and did not the Church live in yearly expectation of the gelieral apostacy which was to precede the e n d ? I should suggest that into the original meaning of the word there flowed a stream of association from a controversy, which was more or less familiar to every educated Christian from boyhood-a controversy ia which the word 'Catholic' suggested (I) truth absolute and admitting of no exception, and ( 2 ) truth which in practice had to protest against yet tolerate error,' 'accedere non indulgere,' as the grammarians say, No doubt the connexion would be much clearer if ' Catholic faith ' and ' Catholic doctrine ' preceded instead of following ' Catholic Church,' as they appear to do. But streams of association do not always folioar logical courses. And it is worth noting that by Origen's time s a 8 o h ~ ~ 6 ~ can be applied to a book in the sense of ' canonical.'" I am inclined to think that much the same may be said of the word 'canon.' I t is at any rate curious that these two words, which have been more than others the watchwords of orthodoxy, bear in grammar practically the same meaning, and are both associated with the analogist doctrine of the universality of grammatical law. The two most prominent ecclesiastical uses of the word are of course (I) Church ordinances, (2) the catalogue, or list of sacred books. T h e first of these followed easily enough from the primary use of ~ a v j f vor a ruler. T h e second is not quite so easy. Souter derives it from the use of ~ . a v j v for a list-quite rightly, I think ; but how did ~ a v j v come to mean a list ! 3 According to Souter the connexion is that the ruler had marks or notches in it, but surely the grammatical use gives a far clearer connexion. For in grammar the word necessarily came to connote a list. The ~ a v t j vwas properly a sentence which laid down a rule, but in practice it was a paradigm.
Perha@ we may add that the ~aBoXinr( were felt to be the safeguards of the heritage of 'EXX?vv~@br, much as catholicity protected the ' faith once delivered.' Cf. Sext. Emp. adz. Gmmm. X. 179. So3ter, Histo'ry of the Canon, p. 179,but he does not give the references. 3 As it is frequently supposed. I think, that this Church use of ' canon ' Is derived from, or a t any rate preceded by, the literary use of the word by the Alexandrine grammarians for the list of the classical authors in any department, I take this opportunity to comment on a doubtful piece of lexicography. T h e Paris Stephanus s.u. navrjv nas the following : ' Similiter ~avijv dicitur catalogus auctorum classicorum, prae ceteris legendorum, a grammaticis Alexandrinis confectus, u. Ruhnken, Htitor. O r . p. xciv.' Evidently following this Liddell and Scott have 2.u.. 'in Alexandrine Grammar collections of the a s being old Greek authors were called ~ov6vrr, models of excellence, classics, Ruhnken, Hist. Or. Crtt. Graec. p. xciv. Cf. Quintilian X . I , 54 ncver n s u r g e s t the word and 59 ' F u t Q ~ ~ i n t i l i e
xaudv, and when I turn to Ruhnken I find that he says nothing of the kind. Not only does he give no evidence, but he never even suggests that such lists were called nav6vrr. H e himself calls them 'canons,' but that is all. A rather different view is given by Westcott, History of tihe Canoit, Appendix A : ' One Instance of the metaphorical use of the word requires special notice. T h e Alexandtine grammarians spoke of the Classic Greek authors as a whole a s b n a v r j v , the absolute standard of pure language, u. Redepenning. a perfect model of composit~on, Ortgimes I. 1 2 . ' Here again we seem to have the same confusion between 7 b ujlpa~vov and rb unparvb~~vov.F o r Redepenning says nothing of the sort. ' H e merely remarks, referring to Quint. X. I, 54, that no living authors ere included by the Alexandrines in ' dem bald allgemein anerkannten Kanon.' T h e fact that Quintilian uses the word ovdo of these lists s:~g; gests that the Greek name was rdtir, whicl~ , confirmed by some of the later evidence q lo.<4 by Ruhnken.

36

F. H. COLSON

W e may say in the form of a sentence that nouns in -a make their genitive in - q ~ but , in practice the schoolboy recited poircra-Fo;crq~, etc., as we actually find in the Theodosii Cnl~oi&cs. T h u s again the educated Christian thought of it as a catalogue, and one moreover which carried with it an authority, to which indeed exccption was sometimes taken, but which did not in analogist theory at least adniit of exception. I t is possible that this disquisition on these two words may seem to some as trivial as it certainly is speculative. And I confess I have an object somewhat outside my main subject. I t seems to me that the terminology of the studies, which formed the staple of general education, deserves more study than it generally receives. I t is not sufficiently remembered that grammar in particular was a science which aroused a vast amount of interest, and that it was, perhaps even more than rhetoric, a study with which the upper and middle classes at any rate were thoroughly familiar. A terminology created under such conditions has a force, a warmth, a colour, which makes its influence felt over fields widely removed from it. Further, I think, we may expect that such an influence would be felt, not so much in what we call the classical literature with its firm traditions, as in that great literature of the middle classes which, untrammelled by tradition, was called into existence by the Church. T h e influence of philosophy on Christian thought and language has of course been examined very carefully. T h e influence of grammar and rhetoric is doubtless of much less importance, but it deserves more careful treatment than, a s far as I can judge, it has at present received.

F. H. COLSON.