tacism. riossus 414 A.U.C., Verrucossus 521 A.U.C. in the period of rhoconsular Fasti), where there was also a nasal to protect the sound from further corruption', which ōsus. lasted on, as an element in the vowel, till quite late times. Labōsus clearly belongs here, whether it is for *labos-ōsus (like clamosus dolōsus fragōsus and others given by Roby) or, as seems at least plausible, formed not from the stem labos-, but from the verb labāre, on the analogy of clamare clamōsus onerare onerōsus criminare criminōsus mõratus mōrōsus dolere dolōsus and others. The meaning 'slippery' would give a more picturesque tinge to the two passages in Lucilius where it occurs (Non. 8. 46), 'iter labosum,' and 'labosas Tantalu' qui poenas ob facta nefantia luvit,' which would describe the cup slipping from his grasp. The commonest case is of course that of participles and perfects in s. After a long vowel or consonant the B. participles s was written single in Quintilian's time, but here etc. in Cicero's the ss was kept in writing, though in pronunciation then, as afterwards, it was probably kept only after short vowels, as in missus, where it was always written from the time of the first introduction of double consonants. If Brugmann's view of the pro-ethnic character of the change of tt to ss in Italy were correct, all participles like fusus would be citable as evidence in support of the theory advocated in this essay, v. supr. § 30, p. 39 foll. The derivation of caussa seems to me so obvious and certain that I can hardly believe it has not been suggested y. caussa quabefore. On inscriptions of the republic (e.g. C. I. L. sillus. I. 198, 556, and 533) the ss appears regular, and in the passage in Quintilian (7. 20) already alluded to (§ 51 (2) ß. n.) it is 1 This n appears not to occur in any inscription (v. Index to C. I. L. 1. ‘n adjecta,' where Mommsen gives only vicensumus etc.), but Ribbeck reads it from the MSS. in two passages of Vergil. 2 v. App. A. The fere in the passage there cited seems to me to imply that the remark applied only to the large class of words, mainly participles, in which s or ss came from tt. 3 For the well-known cases of a long vowel with a single s where from the form of the present we should have expected a short vowel with a double ss (cāsus from cădo, fūsus from fundo), v. Osthoff, Perf. Exc. vi. p. 537. given explicitly as one of the words which were so spelt 'Ciceronis temporibus paulumque infra.' Apart however from spelling, which is at best very unsafe evidence unless independently supported, we have the forms of the compounds incusare, excusare, where the weakening of au to u shews there was no accent on that syllable, from in-causáre, ex-causáre. The s therefore we should expect to have become r if it had been single. These forms however betray the secret by their resemblance to excusum incusum. Is not caussa a past passive participle of an original *caudo, to smite or cut (cf. cauda, 'the smiter,' caudex, 'the bark, or the stump or plank cut off,' also Caudium, the place where the valley divided or opened out into the plain)? Res caussa would = 'res decisa' 'res judicata,' and a glance at the article in any dictionary will shew that the forensic signification of the word is the oldest. A large number of words meaning to 'judge' mean properly to 'cut,' e.g. «pivo cernere decide. The form *caudo must have died out in this derived meaning and cudo have been substituted in the literal sense from the compounds incudo etc., just as spicio1miniscor have banished specio meniscor, and cludo had supplanted claudo in all writers except Cicero at the end of the republic. It had probably only been preserved so long from the influence of includo etc., by such independent forms as claustrum. An incidental advantage of this explanation is to get rid of the irregular u in the present stem of cudo. * In quasillus an original ss has been reduced by the later accent, as in curális from currus. Gr. κάθος shews the root *qath. *qath-lo- became quālo- in pro-ethnic Italic or Latin, *quath-tlo- or a diminutive *quath-tillo- became quassillus in Latin. The same may have happened in Masurius, or the ss which sometimes occurs in the name may be due to the wish to express the accent (§ 46, p. 65 supr.) on a short syllable (§ 27, p. 35 supr.). On the other hand there is no reason for the loss of a second s in casa, nor for supposing an original ss in quaeso, unless indeed the subjunctive meaning 'I will beg' or 'let me beg' which Brugmann supposes, be considered more 1 Given by Br. Gds. § 65, Anm. 2. 2 v. § 30, p. 39 supr. appropriate to the parenthetic 'quaeso' than the simple 'I beg.' Where it was used as a part of the same verb as quaerit quaerimus cet. quarébam quaerémus cet. the r, which was phonetically regular in these, forced itself into quaero also, but where it remained out of system as a mere rhetorical particle (cf. Engl. prithee, please) it kept the phonetic form proper to itself. The passage in Quintilian (1. 6. 33), even as restored by Halm ('satis est vetus quaeso: quid opus est quaesito dicere?'), seems too uncertain to be worth discussing here. The singleness of the s in nasus is shewn by the Sanskrit nasa, O. Bulg. nosu, Engl. nose cet. Pesestas is probably dialectic; in Latin the regular form would have been *pestas like clamōsus for *clamo sōsus. 52. I have now only to give the list of words which constitute the evidence for and against the theory in Latin, with brief comments where they are needed. They are as follows: 1. Words in which s is preserved by accent (i.e. in which it occurs between two vowels after an accented syllable and is not followed by i or u, or, if so, is not preceded by i or u or a long vowel or diphthong). 2. Words in which s has become r after an unaccented syllable. 3. Words in which s has become r after an accented syllable and is not followed by i or u. These have to be explained. 4. Words in which s has become r after an accented syllable under the conditions described. 5. Words in which we have s between vowels after an unaccented syllable. Of these there are only four: Pusillus asellus (infr. § 54) pesestas quasillus v. supr. B. Evidence as to the cause of the change of s to r. 1. List of words in which s is kept by accent. agāso etc. vásum § 11, 54. The words agáso equíso ('stable-boy' 'jockey ') like the proper names Kaéso Píso would be mostly used as appellatives, in the Vocative, and hence its accent would prevail over that of the oblique cases agasónem etc. They seem to be a genuine Latin formation. ásinus. The anaptyptic vowel must have been earlier than 450 B.C. since *asnos would have become *arnos, v. p. 14 supr. The simple word has kept the s in aséllus. caésaries, Caésar ('cum caesarie natus' Fest.) have usually been compared with the Sanskrit kesara (also written keçara) 'hair', and is traditionally distinguished from Kaeso etc. ('a caeso matris utero') which had ss = tt. For the accent of caesaries v. supr. § 47, p. 65 foll. fusus 'a spindle' apparently occurs first in Catullus (Forc.). Is it a past participle, or should we compare Skt. bhush 'to adorn, set off, arrange around'? It is perhaps necessary to remark here that at the time of the rhotacism the vowel of the terminations (domin)-us (milit)-is cet. was ŏ, ě not u, i. immúsulus. Here also the -u(lus) of later Latin was certainly o at the time of the rhotacism. I do not know that any derivation has hitherto been given for this word. That suggested (supr. § 50, p. 70) from mus, is perhaps better than nothing. positus may of course have been still regarded as a compound at the date of rhotacism. It was certainly not so regarded when posui was formed which however is later than Plautus. In compounds it sank to 'postus and this by the side of the regular (re)-postvi1 appears to have given rise to (re)quistvi from *requistum (cf. quaestus quaestor) which finally authorised quaesitum, the original flexion having doubtless been *quaessi *quaestum like gessi gestum. pásus has kept pusillus and then itself dropped out of use. It must however have been a borrowed word originally. Pūsus: puer(us) as Umb. onse: humerus; Osc. Niovμous: numerus; rusem; ruris; Fluusa: floris. quaeso. quaeséndam etc. (Enn. ap. Fest. 258) are analogy forms. rosa. Nothing but desperation could compel an etymologist to regard this word as borrowed. rõsidus. If this form really occurs the vowel of the second syllable was probably intermediate between i and e, and therefore not sufficient to drag back the s tor if it was not predisposed in that direction. This may be the case in luridus ('mala lustra') where the u would have a more directly rhotacistic influence than the ō of rõsidus. siser. It seems more probable that the Greek σíoapov is borrowed from this, than conversely, from (1) the change of declension which is more natural from the consonantal to the o class, (2) the change of the vowel, (3) the irregularity of the σ between vowels in Greek. 1 The s in posici would be kept by that of positus. |