43. Nothing, I say, whereby that which you are comparing with the other may be made more like it or less like it, ought to be brought in from outside, for the purpose of aiding the decision. Therefore, since they do not know in what way the likeness ought to be drawn, they are incompetent to speak about Regularity. I should have said this more plainly, if I were not wishing now to speak more briefly because later on a these matters are to be treated at greater length. Accordingly it is sufficient now to have touched upon them as far as is connected with the general nature of words.
44. I shall next speak of what concerns the individual parts of speech. Since there are several methods of division thereof, I shall now take by preference that by which speech is according to its nature divided into four parts: that which has case-forms, that which has time-forms, that which has neither, that in which both case and time are indicated. Some grammarians call these the parts respectively of naming, saying, supporting, joining: the part of naming is said to be such words as homo ‘man’ and Nestor, that of saying such as scribo ‘I write’ and lego ‘I read,’ that of joining such as scribens ‘writing’ and legens ‘reading,’ that of supporting such as docte ‘learnedly’ and commode ‘suitably.’
45. The kinds of naming are four, of which the words which are like (masc.) quis, (fem.) quae ‘which’ have by certain grammarians been called Provocables; those like scutum ‘shield’ and gladium ‘sword’ have been called Vocables; those like Romulus and Remus have been called Proper Nouns; those like (masc.) hic, (fem.) haec ‘this’ have been called Pronouns. The two middle kinds are called Denominations; the first and last are called Articles. The first class is indefinite, the second is almost indefinite, the third is almost definite, the fourth is definite.
46. Each of these ought to be threefold in nature, as concerns gender, number, and case: gender, whether it is masculine or feminine or neuter, as masc. doctus, fem. docta, neut. doctum ‘learned’; number, whether it denotes one or more, as masc. sing, hic, pl. hi ‘these,’ and fem. sing, haec, pl. hae; case, whether it is in the nominative, as Marcus, or in an oblique case, as Marco, or in a non-distinctive case-form, as ovis ‘sheep.’
47. Now that these divisions have been made, examine them one by one, that you may the more easily see that there are nowhere any Regularities which we ought to follow. To be sure some words had to have three several forms, as in this example: masc. humanus, fem. humana, neut. humanum ‘human’; but some have only two apiece, like cervus ‘stag,’ cerva ‘hind,’ and certain others have but one, like aper ‘boar’; and so on with many others. Therefore Regularity simply does not exist in classes of this sort.
48. In number also, as pater ‘father’ denotes one and patres ‘fathers’ denotes more than one, all ought in this fashion to have two forms. But many are singulars only, like cicer ‘chickpea’ and siser ‘skirret’ — for nobody says cicera and sisera in the plural; and there are words that are plurals only, such as salinae ‘saltworks’ and balneae ‘public baths’: for from these there are no singulars salina and balnea in use. Nor from the singular balneum does general usage make a plural: for because balneum ‘bath’ is like praedium ‘farm estate,’ there ought to be plurals, balnea as well as praedia, but this is not the case. Therefore in these also there is no Regularity.
49. Some have both nominatives and oblique cases, others have nominatives only, others oblique cases only: both, as in nom. Iuno, gen. Iunonis; nominatives only, as in Iupiter, Maspiter; oblique cases only, as in gen. Iovis, dat. Iovi. Therefore in these Regularity does not exist.
50. Now let us look into those of the fourfold division. First, if there were Regularity in the indefinite articles, the proper forms would be feminine quae, gen. quaius, like masculine quis, gen. quoius; and as quoi is dative to masculine quis, so quae would be used as dative to feminine qua. For it is similar by proportion: dea bona qua ‘a good goddess who’ is like deae bonae quae ‘to a good goddess to whom.’ Similarly, the nom. masc. ques stands in the same relation to acc. quos, as the sing. masc. quis to the acc. quern; therefore for the qui homines ‘which men’ which is now used in the nominative, we ought to say ques.
51. Besides, as from masc. is ‘this’ there is dat. ei, so from fem. ea there would be spoken a dat. eae, which is now actually spoken as ei; and like the dat. pl. in ieis viris ‘to these men,’ there would be pronounced a dat. pl. fem. eais mulieribus ‘to these women.’ And as in the nominatives there are masc. is, fem. ea, in the oblique forms there would be masc. eius, fem. eaius; but now eius is said for the genitive not only in the masculine and the feminine alike, but even in the neuter articles, as eius viri ‘of this man,’ eius mulieris ‘of this woman,’ eius pabuli ‘of this fodder,’ although masc. is, fem. ea, neut. id are distinguished in the nominative. I have touched upon this classification more sparingly, because I am of opinion that the copyists will not take proper care in transferring these quite confusing matters.
52. From the appellations which come nearest to the indefinite nature of the articles and are called common nouns, such as homo ‘man’ and equus ‘horse,’ there are four kinds of derivation: one of name-giving, as equile ‘horse-stable’ from equus ‘horse’; the second that of the cases, as accusative equum from equus; the third that of augmentation, as albius ‘whiter’ from album ‘white’; the fourth that of diminution, as cistula ‘little box’ from cista ‘box.’
53. The first class, as I have said, is that in which words in the nominative are derived from some part of speech, as balneator ‘bath-keeper’ from balneae ‘public baths.’ This class has in general three sources, because it develops from a common noun, as venabulum ‘hunting spear’ from venator ‘hunter,’ and from a proper name, as Tiburs ‘man of Tibur’ from Tibur, and from a verb, as cursor ‘runner’ from currere ‘to run.’ In none of these will you see Regularity preserved.
54. First, although from ovis ‘sheep’ and sus ‘swine’ there are said ovile ‘sheepfold’ and suile ‘hog-sty,’ there is no bovile from bos ‘ox’; and although avis ‘bird’ and ovis ‘sheep’ are alike, we do not say oviarium from ovis as we say aviarium’ aviary from avis, nor do we say avile from avis as we say ovile ‘sheepfold’ from ovis; and although there ought to be a sediculum ‘chair’ from sessio’sitting’ like cubiculum ‘sleeping-room ‘from cubatio ‘reclining,’ there is not.
55. Since a shop where wine is sold is called vinaria from vinum ‘wine,’ and cretaria from creta ‘chalk,’ unguentaria from unguentum ‘perfume,’ then if words went in regular fashion a shop where caro ‘meat’ is sold would be called carnaria, one where pelles ‘hides’ are sold would be called pelliaria, one where calcei ‘shoes’ are sold would be called calcearia, instead of laniena ‘butcher’s shop,’ pellesuina ‘leather-shop,’ sutrina ‘cobbler’s shop.’ And just as from unus ‘one’ comes the plural uni ‘one set of,’ and from tres ‘three’ comes trini ‘three each,’ from quattuor ‘four’ quadrini ‘four each,’ so from duo ‘two’ there should be a duini and not a bini ‘two each’; also, after quadrigae ‘team of four’ and trigae ‘team of three,’ there should be rather duigae than bigae ‘team of two.’ There are a great many examples of this class, but I pass them by, since he who has had his attention called to them cannot fail to notice them.
56. If words which develop from proper names ought to be alike if from like names, then since Parma and Roma are identical we shall say Romenses, like Parmenses; or since Roma, Nola, Parma are alike, we shall say Parmani, like Romani and Nolani. And from Pergamum and Ilium we shall have not only Pergamenus ‘Pergamene,’ but also Ilienus; or like Ilius and Ilia ‘Ilian’ male and female, we shall say Pergamus and Pergama respectively for a man and a woman of Pergamum. And since Asia and Libya are like names, we shall call the people Asiatic and Libyatic.
57. The words which are made from verbs are such as scriptor ‘writer’ from scribere ‘to write’ and lector ‘reader’ from legere ‘to read’; that those also do not preserve a likeness can be seen from the fol
lowing: although amator ‘lover’ from amare ‘to love’ and salutator ‘saluter’ from salutare ‘to salute’ are formed in like manner, there is no cantator ‘singer’ from cantare ‘to sing’; and though we say “I am tired with metendo ‘reaping’ and ferendo ‘carrying,’” the words from these do not represent a like relation, since there is no fertor ‘carrier’ made like messor ‘reaper.’ There are likewise many others of this class in which we follow usage rather than conformity to the verbs.
58. Besides these there are other words which also originate from verbs but are unlike those of which we have already spoken, because they have both cases and tenses, whence they are called participles. And as many verbs have opposite forms, such as amo ‘I love,’ amor ‘I am loved,’ lego ‘I read,’ legor ‘I am read,’ from amo and all verbs of this kind there develop present and future participles, such as amans ‘loving’ and amaturus ‘about to love,’ but from these verbs the third form which ought to be made, namely the past participle, cannot be found in the Latin language: therefore there is no Regularity. So also from amor ‘I am loved,’ legor ‘I am read,’ and verbs of this kind the word of this class is made for past time, as amatus ‘loved,’ but from them none is made for the present and the future.
59. Therefore there is no Regularity, especially since such a great number of words has perished in this class which we are mentioning. In these verbs which have not both voices, such as loquor ‘I speak’ and venor ‘I hunt,’ we none the less say loquens ‘speaking’ and venans ‘hunting,’ locuturus ‘about to speak’ and venaturus ‘about to hunt,’ locutus ‘having spoken’ and venatus ‘having hunted.’ This is not according to the Regularities, since we say loquor and venor, not loquo and veno, whence came the forms given above. The Regularities are the less preserved, because some of the verbs which have not both voices, make three participles each, like those which I have named, and other make only two each, such as those which I shall now name: currens ‘running’ and ambulans ‘walking,’ cursurus ‘about to run’ and ambulaturus ‘about to walk’; for the third forms, those of the past, do not exist, as in cursus sum ‘I am run,’ ambulatus sum ‘I am walked.’
60. But Regularity is not preserved even in those which indicate that something is done with greater frequency; for though there is a cantitans ‘repeatedly singing’ from cantare ‘to sing,’ there is no amitans ‘repeatedly loving’ from amare ‘to love,’ and similarly with many others. The situation is the same in the forms of the plural as in those of the singular: though the plural cantitantes is used, seditantes ‘sitting’ is not.
61. Since there is a class of words which they call compositional, saying that they ought not to be grouped in the same category with the simple words of which I have so far spoken, I shall deal separately with these compounds. Since from tibiae ‘pipes’ and canere ‘to play’ the tibicines ‘pipers’ are named, they ask, If we ought to follow the Regularities, why then from cithara ‘lute’ and psalterium ‘psaltery’ and pandura ‘Pan’s strings’ should we not say citharicen ‘lute-player’ and the rest in the same way? If from aedes ‘temple’ and tueri ‘to guard’ the aedi- tumus ‘sacristan’ is named, why from atrium ‘main hall’ and tueri ‘to guard’ is it not atritumus ‘butler’ rather than atriensis? And if from avis capere ‘to catch birds’ the auceps ‘fowler’ is named, they say, from piscis capere ‘to catch fish’ there ought to be a pisciceps ‘fisherman’ named like the auceps.
62. They remark also that establishments where aes ‘copper’ lavatur ’is refined’ are called aerariae ‘smelters’ and not aerelavinae ‘copper-washery’; and places where argentum ‘silver’ foditur ’is mined’ are called argentifodinae ‘silver-mines,’ but that places where ferrum ‘iron’ is mined are not called ferrifodinae; that those who caedunt ‘cut’ lipides ‘stones’ are called lapicidae ‘stone-cutters,’ but that those who cut ligna ‘firewood’ are not called lignicidae; that there is no term argentifex ‘silversmith’ like aurifex ‘goldsmith’; that a person who is not doctus ‘learned’ is called indoctus, but one who is not salsus ‘witty’ is called insulsus. Thus the words which come from this source also, it is easy to see, do not observe Regularity.
63. It remains to consider the problem of the cases, on which the Aristarcheans especially exert their energies.
First, if in these there were Regularity, they say that all names and articles ought to have the same number of cases; but that as things are some have one only, like all individual letters, others have three, like praedium praedii praedio ‘farm,’ others four, like mel mellis melli melle ‘honey,’ others five, like quintus quinti quinto quintum quinte ‘fifth,’ others six, like unus unius uni unum une uno ‘one’; therefore in cases there are no Regularities.
64. Second, in reference to what Crates said as to why those which have only one case-form each are not used in the forms alpha, dat. alphati, gen. alphatos, because they are Greek letters — if the same answer is given to me as to Crates, that they are not our words at all, but utterly foreign words, then I shall ask why the same persons use a full set of case-forms not only for our own personal names, but also for those of the Persians and of the others whom they call barbarians.
65. Wherefore, if these proper names were in a state of Regularity, either they would use them with a single case-form each, like the words of the Phoenicians and the Egyptians, or with several, like those of the Gauls and of the rest: for they say nom. alauda ‘lark,’ gen. alaudas, and similarly other words. But if, as they write, they say that the Greek letters received names with but one case-form each for the reason that they really belong to the Phoenicians, then in this way the Greeks ought to speak our words in six cases each, not in five: inasmuch as they do not do this, there is no Regularity.
66. If Regularity existed, they say, no case ought to be used in two forms; but the opposite is found to occur. For without censure quite commonly some say in the ablative singular ovi ‘sheep’ and avi ‘bird,’ others say ove and ave; in the plural, the nominative is puppis ‘ship’s sterns’ and restis ‘ropes,’ also puppes and restes; likewise there is the fact that in the genitive plural of words of this class there are used the variant forms civitatum ‘of states,’ parentum ‘of parents,’ and civitatium, parentium, and in the accusative plural monies ‘mountains,’ fontes ‘springs,’ and montis, fontis.
67. Likewise they say that, if there is Regularity, like forms ought to be made from like words declined alike, and that this can be shown not to take place; that therefore this theory is to be rejected. And yet this failure can be shown; for how can anything be more alike than gens ‘clan,’ mens ‘mind,’ dens ‘tooth’? Despite which their genitives and accusatives in the plural are unlike; for from the first word are made gentium and gentis, with I in both, from the second come mentium and mentes, with I in the former only, from the third dentum and denies, with I in neither form.
68. So likewise since surus ‘stake,’ lupus ‘wolf,’ lepus ‘hare’ are alike in the nominative, they ask why there is not said in like fashion suro, lupo, lepo. But if the answer is given that they are not alike, because in the vocative we use the unlike forms sure, lupe, lepus (this to be sure is what Aristarchus wished to say in reply to Crates; for when Crates had written that Philomedes, Heraclides, Melicertes were alike, he said that they were not alike, because in the vocative Philomedes is said with a short E, Heraclide with a long E, Melicerta with a short A), in this, they say, Aristarchus did not realize that he was not giving a solution of the question that was asked.
69. For in this way, whenever there was any difference in the oblique eases, he could say that for this reason the nominatives were not alike; but since the question is whether the two nominatives are like each other, or not, there should be nothing brought in from outside, they say, as to why they are alike or different.
70. Likewise if there were Regularity, then just as they say in like fashion the nominatives aves ‘birds,’ oves ‘sheep,’ sues ‘swine,’ they would say in the geni
tive suium just as they do avium and avium. If there is Regularity, he says, why do the people say dei ‘gods’ in Dei Penates ‘Household Gods’ and Dei Consentes ‘United Gods,’ although in the nominative singular deus ‘god’ is just like reus ‘defendant,’ ferreus ‘of iron,’ and so the plurals should be reei, ferreei, deei?
71. Likewise they ask, if Regularity exists, why do people all say the Temple Deum Consentium ‘of the United Gods’ and not Deorum Consentium? Likewise, why do they say a thousand denarium ‘of denarii’ and not a thousand denariorum? For this word denarius is in form exactly like Vatinius and Manilius, and therefore denariorum ought to be used as genitive, like Vatiniorum and Maniliorum. They say also that a cavalryman’s horse is worth not a thousand assarium ‘of assarii,’ but a thousand assariorum; for from the singular assarius comes the plural assarii, and from that should come the genitive assariorum.
72. Likewise, according to their theory, it is with long second syllables that we should pronounce Hectōrem and Nestōrem; for the accusatives quaestorem praetorem Nestōrem, and the genitives quaestōris praetōris Nestōris correspond to the nominatives quaestor praetor Nestor. And we ought not to say quibus das ‘to whom you give,’ his das ‘to these you give’; for it is dative his and quis, like nominative hi and qui, or else it is hibus like quibus.
73. Although the dative patri familias ‘to the father of the household’ is used, still, if they wished to follow Regularities, they ought never to have said nominative pater familias, because the word is genitive familiae, like Atiniae and Catiniae, and therefore nominative familia, like Atinia and Catinia. Likewise, they ought not to say patres familias ‘fathers of a household,’ but as Sisenna writes, they should say patres familiarum ‘fathers of households.’
74. Nor ought usage to fluctuate, in that some said herds boum ‘of cattle,’ others boverum, and others said statues Ioum ‘of Jupiters,’ others Ioverum, since like the genitive Iovis bovis struis were the accusative Iovem bovem struem and the dative Iovi bovi strui; and since these agreed in the oblique cases, they ought not to have varied in the nominative forms, in which it is now customary to use different formations, Iupiter for Ious, bos for bous, strues ‘heap of offering-cakes’ for strus.
Delphi Complete Works of Varro Page 54