Published online by Cambridge University Press: 22 April 2026
This study proposes to examine the phonetic ongm of the Latin diminutive suffixes in -ellus -a -um and -illus -a -um, and to trace the lines of their analogical extension as independent formative elements, a phenomenon which is later than the origin of -ello/ā and -illo/ā- through phonetic processes. Since scholars have expressed several opinions on the phonetic origin of these suffixes, it is perhaps justifiable to examine the validity of the view most commonly held, by adducing all the extant pertinent materials, inasmuch as no such detailed presentation has yet, to my knowledge, been made. Moreover, little or no consideration has been given to the problem of the extension of the suffixes -ello/ā- and -illo/ā- from words or groups of words where their development was phonetic, and of the possible points of departure of this extension.
page 10 note 1. The notations -ello/ā- and -illo/ā- will hereafter be used, for the sake of convenience, to indicate the suffixes -ellus -a -um and -illus -a -um.
page 10 note 2. Whether the suffix proper is in a given case -ello- -illo- or -cello- -cillo- is irrelevant to this study, since the sound c which often commences the diminutive suffixes (and is limited almost exclusively, at least during the Republican period, to words of the third, fourth, and fifth declensions) does not affect their vocalism. The diminutive suffix -culo/ā-, upon which are built -cello/ā- and, by analogy, -cillo/ā-, is derived from -k-elo/ā-, this in turn from -qo-, a diminutive formans obsolete in Italic (cf. homun-c-iō), plus the secondary or diminutive -elo/ā-; see Leumann 172.IV.A.3.
page 10 note 3. For the purposes of this study, I have considered the authors from Livius Andronicus to and iucluding Vitruvius, Vergil, Horace, and Ovid as Republican.
page 11 note 4. The suffix -ello/ā-, as seen in porcellus (to poreus porculus, from *porc-elo-s) is conveniently termed ‘double’, as distinct from the ‘single’ suffix of fäbella (to fābula).
page 11 note 5. In collecting both the Republican and the Imperial examples which have been preserved by manuscript tradition and on epigraphical remains, I have utilized the materials gathered by Otto Gradenwitz in Laterculi Vocum Latinarum and the lists of C. Paucker, KZ 23.169-78.
page 11 note 6. The lack of diminutive meaning in Latin nouns of diminutive formation is a subject treated by J. Friedrich, Deminutivbildungen mit nicht deminutiver Bedeutung, 43-75; see also F. Conrad, Die Deminutiva im Altlatein, Glotta 19.127-148 (1931) and 20.74-84. Conrad considered (134) ' bei welchen Worten der Dichter (Plautus) durch die suffixale Form zweifellos ausdrücklich den Begriff der Kleinheit hervorheben wollte. Nach, sorgfältiger Prüfung bin ich zu dem Ergebnis gelangt, dass dieser Fall bei keinem der plautinischen Deminutiva zutrifft.' He concluded (144) : ' Das Deminutivsuffix ist das Zeichen eines Affekts, mag dieser auch noch so geringfügig sein.'
page 11 note 7. See Brugmann, Gdr.2 2.1.365-7, and Leumann 172.IV.A and B; words ending in -ulo/ā- which have anaptyctic u before l also form diminutives in -ello/ā- and -illo/ā-, as e. g. tabula (tabella) and crepitāculum (crepitācillum). Adjectival -elois seen perhaps in vitulus (vitellus), and in the formations pūllus ‘ pure ‘ < *pūr-elos and suillus <*suīn-elos; primary -elo- is seen in oculus (ocellus) and tegulum (tegillum).
page 13 note 1. GL 2.109-12.
page 13 note 2. Consistently with this, Priscian cited in this connection the forms columnella and scamnellum = classical columella and scamellum; since the latter two words owe their -ll- to the assimilation of -nl-, the operation of this principle could not have been known in Priscian's time, when the new formations columnella and scamnellum were made.
page 14 note 3. De deminutivis graecis et latinis liber, 39-43 and 96-103.
page 14 note 4. ' Congenial ' renders the 'maxime congruens ', ' gratior ', or 'genehm ' of the earlier investigators.
page 15 note 5. Cicero, Orator 45.153, had thought āla a contraction of axilla !
page 15 note 6. Schwabe (101) thought the term justified because of the appearance of a sibilant in OHG ahsa, ahsala, Gr.
ξων, Lat. axis.
page 15 note 7. De linguae latinae deminutivis, 3-9.
page 15 note 8. P. W. Corssen, Ueber Aussprache, Vokalismus und Betonung der lateinischen Sprache 1l.250 (1858).
page 15 note 9. Corssen, op. cit., 21.12 (1859); 12.641 (1868). Cf. also Corssen, Kritische Beiträge zur lateinischen Formenlehre (1863), 207 and 424.
page 16 note 10. H. Osthoff, Forschungen im Gebiete der indogermanischen nominalen Stammbildung, 1.17-8.
page 16 note 11. 18.
page 16 note 12. This word Osthoff etymologized, after Ascoli, Le figure italiche 25 : *fā-blulā (primitive *fāblā) > *fā-bel(u)lā > fā-bel'la.
page 16 note 13. Die Deminutive mit doppeltem 1, KZ 23.169-88 (1877).
page 17 note 14. 170.
page 17 note 15. The first of these words is a Late Latin by-form of cēpula with gemination of l, the second perhaps an etymologizing of Caesulla, feminine of Caesō; see Hofmann s. vv. cēpe and caesius.
page 17 note 16. See Leumann 144 e. α; Paucker overlooked the now accepted view of Osthoff, Forschungen im Gebiete der indogermanischen nominalen Stammbildung, 1.190 sqq.
page 17 note 17. De deminutivis graecis et latinis liber, 96.
page 17 note 18. 171.
page 18 note 19. Genuswechsel der Deminutive, ALL 4.169-88(1887).
page 18 note 20. 174.
page 18 note 21. 172.
page 18 note 22. See Brugmann, Gdr. 11.201 and 237 (1886) for these notations, and 11.192 note for a bibliography of the earlier articles on this subject.
page 18 note 23. 172.
page 18 note 24. In the first place, the vowel (regularly e for diminutives of -ro- stems) developed before the assimilation of -rl- to -ll-, and remained the same after assimilation took place; and secondly, although e or i is the preponderant vowel before -ll-, the theory of the ‘ congenial vowel ‘ is inadequate to explain the phenomenon, if only because of the existence of such words as hotnullus and satullus. In the present instance, the difference of the ending of Lrānslillum from that of all other diminutives from -ro- stems would justify one in thinking of an extra-phonetic process; see Sommer2 82.
page 19 note 25. Grundriss der vergleichenden Grammatik der indogermanischen Sprachen, 21.1.193 (1889).
page 19 note 26. Brugmann recognized the original suffix as -elo-, Gdr.2 2.1.366-7 (1906).
page 19 note 27. Griechische und Lateinische Sprachwissenschaft, ed. 2 (1890), 2.271 and 285.
page 19 note 28. Stolz recognized (285) that the ṛ from which Latin er developed is PIt. sonant or syllabic ṛ, and not the ṛ of PIE.
page 19 note 29. Stolz left it doubtful (310) how he would have explained the ending of axilla, where the phonetic group ḷl is involved.
page 19 note 30. Historische Grammatik der Lateinischen Sprache, 1.135 and 205.
page 20 note 31. Stolz referred ocellus to the late syncopated form oclus of the Appendix Probi, in order to bolster up his theory that diminutives from primitives in -ulo- passed through a -ḷlo- stage !
page 20 note 32. 135; strangely enough, Stolz considered scabellum (to scamnum), and not scabillum, as the phonetic form.
page 20 note 33. The Latin Language, Oxford 1894, 333.
page 20 note 34. Nor in 1915, when he reprinted his old views in the second edition of his Short Historical Latin Grammar, 188.
page 20 note 35. See Latin Language, 273 and 278 (the PIE sonant nasals and liquids), where Lindsay tacitly passed over PIt. ṇ and ḷ, and granted PIt. ṛ (> Latin -er-) only dubious existence.
page 20 note 36. 180.
page 20 note 37. According to Lindsay (232), unaccented and accented -rǐ- became -ěr- when preceded by a consonant, as in incěrtus < *íncrǐtos, těr < *těrs < *trǐs, and těstis < * těrstis < *trǐstis.
page 20 note 38. Ӗ und Ĭ im Lateinischen, 58-69.
page 21 note 39. The sound-group -ell- with PIE 7#x04D7; was again treated by Niedermann, IF 15.111-3, where he maintained his views in answer to a review of his dissertation by Stolz, IF 13.99-100. Stolz disallowed syncope in the third syllable, which was necessary if Niedermann's groups -sn(o)l-, -er(o)l-, and -el(o)l-, were to become Latin -ell-, and would derive asellus from *asn(o)-lo-s (to a primitive *asnos) > *asṇlos; this form, as Niedermann rightly observed, could not become asellus but only *ānulus. Similarly, Stolz took porcellus from *porcḷlos < *porc-lo-lo-s, while Niedermann, taking Lithuanian parszēlis ‘*porcelos porculus ‘ as his basis, posited *porcelo-lo-s as the starting-form.
page 21 note 40. Niedermann's own formulation, Ӗ und Ĭ 69, reads : ‘Indogermanisches ě blieb vor -ll- unverändert, dagegen unterlag durch auf italischem Boden silbenbildend gewordene Liquida oder Nasalis erzeugtes ě vor -ll- dem Wandel zu ǐ. ‘ Cf. also IF 15.111.
page 21 note 41. So too (62), bacillum from *bacḷlom, but incorrectly tegillum from *tegḷlom (the suffix of the primitive shows primary -elo-); tegillum certainly has -illum by extension, a phenomenon not investigated by Niedermann, as lying beyond the scope of his dissertation.
page 21 note 42. This word Niedermann incorrectly referred to an unattested primitive *pistrum, disallowing, IF 15.113 note 1, the phonetic possibility that pistillum was diminutive to pīlum; see Sommer, Kritische Erläuterungen 23, and Leumann 144. e. α.
page 22 note 43. Niedermann's attempt (69) to show that i arose subsequent to assimilation, and not prior to it, must be judged unsuccessful (he compared the medial e of dēcerno <* dēcřnō <*dēcrǐnō with the medial i before rr in equirria, which he derived from *équocṛsiā < *equocersiā; but see Hofmann s. v. currō). If Latin -ll- effected the change from e to i in some cases, it would have effected the same change in all cases, unless hampered by a series of analogical interventions.
page 22 note 44. Handbuch der lateinischen Laut- und Formenlehre, ed. 1, 55-7.
page 22 note 45. The ‘double’ diminutive suffix in -illo- was not considered by Sommer.
page 22 note 46. Useless because the original diminutive suffix was now recognized to be -elo-, not -lo-. Sommer was perhaps mystified by the ‘doubleness’ of the suffix, and preferred to regard it as the result, not of conscious duplication of the formans, but of an extra-phonetic process. Or perhaps he regarded syncope in the third syllabic as unlikely; cf. IF 11.227, where Sommer derived lībertās from *louf(e)rŏtāts (becoming *loufertāts by analogy to the nominative * loufer),and not from *loufěr(⊖)tāts.
page 22 note 47. Gdr.2 2.1.366 note 2.
page 22 note 48. Brugmann, Gdr.2 2.1.366-7, established by IE cognation the fact which Niedermann, Ĕ und Ĭ 62-3, had arrived at a posteriori, viz. that the original diminutive suffix was -ĕlŏ-.
page 23 note 49. Another possible objection to Brugmann's phonetics is this : supposing *signolo-m to have become *signolom > *sigṇlom, the product of ṇ, -en-, would automatically become -en- before the following sound and the resulting diminutive could be only *sigellum.
page 23 note 50. Gdr.2 1. xlii.
page 23 note 51. Brugmann thought that the i of simplex (to *sem- ‘one‘), as over against the e of templum, was conditioned by preceding s; but see Hoffmann, BB 18.158; Niedermann, Ĕ und Ĭ 8-25; Sommer2 § 53. I. b.
page 23 note 52. Lateinische Grammatik, ed. 4, 100 and 104.
page 23 note 53. Cf. Stolz, IF 13.97, and see above, note 39. Götze, IF 41.84, rightly observed that the difference in the products asellus sigillum and catellus põcillum necessitates postulation of syncope in the third syllable for -ello-, and syncope in the second syllable for -illo-, respectively; for syncope in the third syllable in general, see Leumann 78.2.
page 23 note 54. This primitive, according to his view, acquired an ‘anaptyctic ‘ i and became Latin geminus, as *asnos became Latin asinus; cf. his Historische Grammatik 1.205, where he accepted the old view of G. Meyer, IF 1.319-20.
page 24 note 55. Under this assumption, porcellus would have to be derived from *porclo-los (to a demonstrably false antecedent form *porc-los), becoming *porcḷlos.
page 24 note 56. Dominance et résistance dans la phonétique latine, 153-67.
page 24 note 57. 155.
page 24 note 58. 156.
page 24 note 59. 157.
page 24 note 60. 167.
page 25 note 61. Juret's later formulations, Manuel de phonétique latine 145-7, are clear, but are not illustrated by adequate phonetic progressions. Apparently, the following would represent Juret's views on the development of diminutives in -ello/ā- and -illo/ā- from stems in -ro/ā-, -no-, and -lo/ā- (bracketed forms not indicated by Jurct) :
*rūtrolom (or [*rūtrilom]) > [*rūtorlom (*rūtirlom)] > [*rūterlom], rūtellum;
*pugnolos > [*pugonlos] > [*pugollos] > pugillus (o > i by palatalization);
*putslolos > [*pussollos] > pusillus (o > i by palatalization).
Cf. Niedermann, Berliner Philologische Wochenschrift 42.294 and 982 (1922), who in a review of the Manuel similarly interpreted Juret's formulations.
page 25 note 62. = Geschichte der indogermanischen Sprachwissenschaft, edited by Streitberg, 2.1.169-70.
page 25 note 63. Thus, for example, Walde accepted Summer's analogical proportion in accounting for the ‘ double ‘ suffix in -ello-, and Brugmann's view that PIt. ṇ became -en-, except after g, where it became ṇ > -eṇ- > -iṇ- > -in-.
page 25 note 64. There has been some hesitation in accepting -in- as the product of PIt. ṇ and -il(l)- as the product of PIt. ḷ(l), in view of the marked divergence of these products from those of PIE ṇ (> Latin -en-) and PIE ḷ (> Latin -ol-); R. G. Kent expresses a dissenting view in The Sounds of Latin, § 124.
page 25 note 65. See Sommer 2 82-3; Leumann, Lateinische Grammatik, Part 1, §§ 102, 104, and 172. IV. A; and Niedermann, Précis de Phonétique Historique du Latin2 (Paris 1931), 51-2.
page 26 note 1. If ager is the primitive of agellus, i. e., is the direct morphological antecedent on which the diminutive is made, then cistula may perhaps justifiably be termed the primitive with respect to cistella, since the latter form is morphologically made on cistula, and not on cista.
page 26 note 2. The terms ' single ' and ' double ' are used throughout chiefly in a morphological sense. Others have treated the ' single ', the ' double ', and the ' triple ' suffix from the semantic standpoint. See Cooper, Word Formation in the Roman Sermo Plebeius 168; W. A. Baehrens, Sprachlicher Kommentar z. vulgärlat. Appendix Probi 121-2; and F. Conrad, Die Deminutiva im Altlatein, dotta 19.140-1 (1931).
page 27 note 3. *Novolus probably never existed; if it did, it is strange that we nowhere find it attested.
page 27 note 4. Anguilla ' eel ' is of course non-diminutive in meaning. The diminutive formation was sometimes employed merely to express a certain relationship in meaning between two words, as in the well-known examples ōsculum ' kiss ', to ōs ' mouth ', and capillus ' hair ', probably to caput ' head '. This function of the diminutive formation may perhaps, for want of a better name, be termed the ' relational'; cf. Friedrich, Demin. Bild. 63.
page 27 note 5. The first attestation of a given word is sometimes an inadequate criterion of its age. All diminutives occurring for the first time in Festus or Paulus Diaconus may justifiably be regarded as early, since the text of these epitomators rests on that of the Augustan encyclopaedist Verrius Flaccus, who had concerned himself largely with archaic terms bearing on Roman religion and history. Generally speaking, diminutives of Imperial date have been listed with the earlier words when the internal evidence of their derivation and their contexts has justified such inclusion.
page 28 note 1. By-form : gamella Scaurus (time of Hadrian).
page 28 note 2. This stem was weakened from -ărā-, since camera (cammara App. Probi) was borrowed from Gr. χαμάρᾱ ‘ gondola ‘. It should be noted that -ā- does not appear in derivatives of -ā- stems, except in fabāginus (to faba), oleāgin(e)us (to olea), and capreāginus (to caprea), where perhaps the whole ending -āgin(e)us is analogical in origin; see Leumann 172. I. B.
page 28 note 3. By-form : cancella -ae CGL 4. 531. 46 (fem, after caulae).
page 28 note 4. Cancer ‘ grating ‘ is found only in Paul.-Fest. 46. 2 M : cancri dicebantur ab antiquis qui nunc per diminutionem cancelli.
page 28 note 5. By-form : canistellus Not.. Tir. 108.96 S, to masc. canister.
page 28 note 6. Paul.-Fest. 45.19 M cana dicunt Graeci, nos canistra et per diminutionem canistella. On the inclusion with the earlier words of diminutives first attested in Festusor Paulus, see § 23 supra, note 5.
page 28 note 7. Masc. capellus Prisc., to caper.
page 28 note 8. By-forms : cletellae Greg. Tur. (clec- cod. B 1), clat- CGL 2.522.18 (influenced by strātum ‘ saddle ‘ ?); and with dissimilation crit- Veget., cret- CGL 5.594. 34, and sing, cratella Chiron. For the meaning of clītellae, see Fröhde, BB 17.303, and Meringer, IF 16.120.
page 29 note 9. Spelled both gravastellus and ravistellus in Paulus Diaconus (-ist- induced by other diminutives in -istello/ā-, such as canistellutn); see Paul.-Fest. 96.13 and 272.7M, and Walde 2 s.vv. grāvastellus and rāvus ‘ gray ‘. For gr- becoming r-, see Leumann 117. For the suffix -aster (-astro-) of unknown origin and pejorative meaning, see Leumann 172. VII.B. See also List 18.41 gravescella.
page 29 note 10. CGL2.470.22 mostellum φάσμα; cf. CGL 2.587.42 monstellum, and 3.457.13 muscellum φάσμα (for u cf. the doublets clōstellum clūstellum; see List 11. 3 and note). On c for t before the palatal vowel e in mūscellum, see Leumann 111 .b and d. For the weakening of n before s and the writing s for ns in mōstellum, see Leumann 105; cf. CIL I2.8 cosol and cesor, and CIL I2.581 cosoleretur (beside consolvervnt).
page 29 note 11. The fem, subst. nigella -ae (CGL 3.595.5 sinonus siriacus) may be a late translation of Gr. μελάνθιον ' Nigella sativa, fennel flower '; cf. Meyer-Lübke REW 2 no. 5915 (' black cumin').
page 29 note 12. Patera, rather than patina, is to be assigned as the primitive, since only the former word has the meaning ‘libation-bowl, offering-dish ‘ in common with patella.
page 29 note 13. Var. plaustellum, to plaustrum; see Sommer 2 78-9.
page 29 note 14. Since puella meant ' girl ' and not specifically ' little girl ', it was perhaps made on puer (puero-) to differentiate the fem, gender in this originally epicene noun. Puera, though attested for early Latin (cf. Suet. Calig. 8 quod antiqui etiam puellas pueras dictitarent), is rare and may perhaps be reckoned as an analogical formation.
page 30 note 15. Puellus is attested principally by the grammarians and was probably never in common usage. It may have been made as a collateral form to puella.
page 30 note 16. Var. spelling pulcellus, to piacer; see Terent. Scaur., GL 7.20.4, and Sommer2 200-1.
page 30 note 17. This diminutive does not take its gender from rāstrum, but from the heterogeneous plural rāstrī (more common than rāstra); for the gender of diminutives in general, see Brugmann, IF 19.215, and A. Weinhold, ALL 4.169-88 (Genuswechsel der Deminutive).
page 30 note 18. By-form scalpellus Celsus, to masc. scalper (Celsus).
page 30 note 19. New formations : scutrilla Thes. Nov. Latin. 567 (Mai), and scutrillus Pompeius Gram. Scŭtella acquired the back-formation scŭtula ' small dish ' Cato, perhaps because of the resemblance of a platter to a small shield (scūtulum). But scūta Lucil. (223 Marx
scūtam Ligneolam in cerebro infixit) is probably a back-formation of scŭtula ' cylinder, barrel ', a borrowing of Gr.
' cudgel '. See Niedermann, Anzeiger f. indoger. Sprach- u. Altertumskunde 19.35, and ALL 14.436.
page 30 note 20. LL 5.174 sembella quod libellae dimidium, quod semis assis. Sēmbella is derived, with syncope and dissimilatory shortening, from *sēm(i)bella < *sēmi-(lī)bella see Walde 2 s. v.
page 30 note 21. See Walde 2 s.v. sternō; the simple monosyllable *stēr ‘ star ‘ was amplified, in Greek by a prothetic
-, in Latin by a diminutive formation. The primitive *stēr, being a consonant stem, perhaps added the formans without a connecting vowel : *stēr-lā > stēlla; see Leumann 172. IV. A. Meillet does not consider stēlla a diminutive, and derives it from *stēl-nā (*stēl- being the doublet form of PIE *stēr in Gr.
); see Ernout-Meillet, Diet. Étym. s. v. stella.
page 30 note 22. Var. reading trāstillum (trastillis Vitr. 5. 12. 3, ed. Rose).
page 30 note 23. Forms such as fenestrula Apul., miserulus Liv. Andr., mitrula Solin., operula Apul., puerulus Cic., and tesserula Lucil. (tessella Sen.) are analogical formations made directly on the primitives fenestra, miser, mitra, opera, puer, and tessera.
page 31 note 24. See note 23.
page 31 note 25. The conditions of syncope in Latin have not yet been fully determined. One may, however, accept as established fact the general principle that syncope in the second syllable was common, whereas syncope in tbe third syllable was uncommon. Götze, IF 41 .80 sqq., considers syncopations of the latter type as phonetic and as the earliest, but adduces many examples which give the impression rather of analogical development; for criticism of his views, see Leumann 78.2. It is uncertain, for example, whether lībertās is to be derived from *loifer(o)tats, or from *loif(e)rotāts becoming *loifertāts, lībertās, by analogy to the nom. sing, of the adj. *loifer(s), līber (so Sommer, IF 11.227). In this study, syncope in the second syllable will be assumed for the development of the following types of diminutives : agellus < *agr(e)los, sigillum < *sign(e)lom, axilla < *agsl(e)lā, and bacillum < *bakl(e)lom; syncope in the third syllable will be assumed for the types : patella < *pater(e)lā, asellus < *asen(e)los, cistella < *cistel(e)lā, and pistrīlla < *pistrīn(e)lā.
page 31 note 26. The original diminutive suffix was normally -ělŏ/ā-, rather than -lŏ/ā- or -ŏlŏ/ā-; cf. the ‘double’ suffix -ello/ā- from -ĕl-ě/⊖/a-, Umbr. katel ' catulus ', Umbr. struçla ' *struiculam ' (ç showing that ē has been syncopated), Osc. famel ' famul, famulus ', Fal. arcentelom ' *argentulum '; OHG farheli (= NHG Ferkel); Lith. parszělis ' porculus '. For example, *porc-eḷo-s > *porc-olo-s (ĕ becoming ŏ before velar l) > porculus; *porc-el-(e)lo-s > *porc-ello-s (ĕ remaining before palatal ll) > porcellus. If porcellus be taken from *porc-ol-(p)lo-s > *porcollos, the change to -ell- must be assigned to the influence of palatal ll; but cf. ampulla <*ampor(e)lā, homullus < *hemon(e)los, and satullus < *satur(e)los, where the influence of ll was apparently in abeyance. See Leumann 102.1 end, and 172. IV. A; von Planta, Grammatik der osk.-umbr. Dialekte, 1.261 and 362, 2.25-6.
page 31 note 27. The phonetic process by which a postconsonantal
,
, r, l, or n becomes through syncope i, u, ṛ, ḷ, or ṇ is termed samprasāraṇa, an expression of Hindu grammarians. That samprasāraṇa, and not merely metathesis, was present in the development of *agr-elos to *agerlos, agellus, is shown by the development of *pōiclelom to pōicillum. If it be assumed with Juret, Dominance et résistance 167, that metathesis was operative in the change of *pōclĭ-lom to pōcillum, one is driven to the unwarranted conclusion that pōcillum was formed in a period when a short vowel in an open medial syllable was pronounced ī, whereas agellus was formed in an earlier period when a short vowel in the same position was pronounced ē. — PIt. ṛ, arising by samprasāraṇa, regularly gave -er- : *ágros > *ágrs > *agerz > *agerr > ager; *sdcr(o)-dōts > *sácṛdōts > sacerdōs. See Sommer 2 § 67; Leumann 102.
page 31 note 28. In other Italic dialects this process produced similar results; see von Planta, Grammatik der osk.-umbr. Dialekte, 1.323-4. Cf. Osc. Aderl[ú] ' Atella ' < *Ātr(o)lā; r perhaps = ṛ in Sab. Atrno ' Aternus ' < *Atr(o)no-, and in Osc. Tantrnnaiúm. Abella was taken by Corssen, KZ 2.17, from *Apr(p)lā (to aper), and by von Planta, op. cit. 1. 336, from *Ag
nolā (to agnus), but is now derived by Hofmann s. v., from *ablonā, ' Benennung der Stadt nach der Apfelzucht '.
page 32 note 29. Capella and lābellum ‘ tub ‘ as old agricultural words, patella and sacellum as ritualistic terms.
page 32 note 30. See above, note 14.
page 32 note 31. See F. Conrad, Glotta 19.135; Conrad, following J. B. Hofmann, Lateinische Umgangssprache (Heidelberg 1926) 140, proposes to express the distinction between the emotional and the intellectual phrase or statement by the term ‘ lust-oder unlustbetonte Begriffe ‘. For the emotional coloring of diminutives, see Wölfflin, Philologus 34.153 and ALL 9.10-1; and Wilhelm Kroll, C. Valerius Catullus (Leipzig 1923), notes on 61.181, 63.35, and 64.103.
page 32 note 32. This word was of course used in a technical non-emotional meaning by medical writers.
page 32 note 33. Cf. Varro ap. Non. 456.9 M oculis suppaetulis nigelli pupuli, which conveys no more to the understanding than would oculis paetis nigri pupuli. The first phrase has emotional coloring, the second is an intellectual statement of exactly the same sense but bare of adventitious suggestion.
page 32 note 34. See § 22 supra, note 4.
page 32 note 35. On the relation of flābellum and flābra, see Friedrich, Demin. Bild. 68.
page 32 note 36. See Thes. Ling. Lat., s.v. castrum.
page 37 note 37. Cf. Varro LL 5.174 nummi denarii decuma libella, quod libram pondo as valebat, et erat ex argento parva.
page 37 note 38. Cf. Suet. Caes. 56 epistulae, quas primus videtur ad paginas et formam memorialis libelli convertisse.
page 37 note 39. So Hofmann, s. v. (1.) cancer, after Skutsch, BB 22.127, who proposed that cancer ‘ bar ‘ is a dissimilated form of carcer(ēs).
page 37 note 40. See Thes. Ling. Lat., s. v. cancellus. Cancellī seems never to have shared the meaning ‘ racing-stalls ‘ with its primitive (?) carcerēs (cf. perh. Cic. Quinct. 36 si extra hos cancellos egredi conabor, where cancellos is equivalent to fines).
page 37 note 41. Vitr. 5.12.3; see Sommer 2 82, and Juret, Dominance et résistance 158, note 1. Juret would derive trānstillum from *trānstrolom, becoming *trānstlolom by dissimilation, the group -lolo- regularly becoming (157) -illo-. But it is difficult to understand why we should not then also have *rāstillus and *rūtillum, to rāstrum and rūtrum, where also the conditions for dissimilation were present.
page 37 note 42. For tigillum with the meaning ‘ small beam ‘, see Liv. 1.26.13, Catul. 67. 39, and Juv. 7.46.
page 34 note 1. Fem. asella Ovid, to asina.
page 34 note 2. Asellus in the meaning ‘ sea-fish (cod ?) ’ may be a translation of Gr.
of the same meaning; see Friedrich, Demin. Bild. 58.
page 34 note 3. Weakened from -ono- in the original form *asonos < *osonos; cf. Gr.
from PGr. *οhονος > *hο-ονος, i. e. ό
! See Hofmann s. v. asinus.
page 34 note 4. Dvenos (CIL I2.4) became OLat. duonos ‘bonus’ (dvonoro CIL I2.9) by vowel-assimilation; but bene < *beē < *
.
page 34 note 5. Var. MSS reading columnella. Columellus (sc. dēns’) ‘eye-tooth’, attested by Chiron and Isidore, is a back-formation from columellāris (dēns) found in Varro; see Sofer, Isid. 128.
page 34 note 6. Columna is perhaps best taken as a pres, medio-passive part, from the root *qel- ‘ project’ and derived from *qelom(e)nā; see Hofmann s. v., where the several views are criticized. Some are of the opinion that Latin -mno/ā- represents a different ablaut grade from -meno/ā-, and is not a syncopated form of the latter (under this assumption, damnum, for example, is from *da-?nno-m, not from *da-meno-m) ; see Meillet, MSL 21.196, and Leumann 172.IX.G. Accordingly, the diminutive columella may be derived in two ways : (1) with syncope and assimilation, from *kelornen- (1e)lā; or (2) with samprasāraṇa and assimilation, from *kelomn(e)lā >- *kolom
lā > *koluminlā]> *columilla, becoming columella by analogy to diminutives in -ello/ā - from -(e)ro/ā- or -eno/ā- stems. I incline to adopt the first derivation; the second is that of Leumann 106; see also Sommer2 83 (indecisive). If regular phonetic development is waived, columella may be explained as made after the model of scamellum (: phonetic scabillum)) to scamnum; thus scamnum : scamellum = columna : columella.
page 34 note 7. Fēmina is from PIE *dhē-menā, cf. Gr. θημένη ‘ giving suck ’; see Walde2 s. v.
page 34 note 8. By-forms : fiscellus Colum. (gender after fiscus), whence the back-formation fisclum Isid., and fiscilla Edict. Imp. Diod.; cf. also Paul.-Fest. 90.2 M fiscellus (-eliō ?) casei mollis appetitor, ut catillones catillorum liguritores.
page 34 note 9. On the Lat. suffix -ino/ā- as seen, e. g., in fiscina, f acinus, pāgina, and sarcina, see Brugmann, Gdr.2 2.1.260 and 267, Leumann 172. IX. A. 4.
page 35 note 10. Var. MSS reading lammella, to lammina; on the syncopated primitive lamna was made lamnula Tertul.
page 35 note 11. Lāmina is perhaps from *stlāmenā ‘ the extended (thing) to (st)lātus ‘ broad ‘; see Leumann 172.IX.G.
page 35 note 12. By-form numellus Colum. For the meaning, cf. Paul.-Fest. 173.18 M numella genus vinculi, quo quadrupedes deligantur, solet autem ea fieri nervo, aut corio crudo bovis, ut plurimum.
page 35 note 13. So Walde2 s. v. numella, after Fay, AJP 25.372. Muller, Altit. Wörterb. 291, derives the word, alter Lidén, KZ 41.398, from *nomnā ‘band ’ < *nop-nā, to napurae ‘straw ropes ’; but *nopn(e)lā would have become *nubilla (?), and then numella only by analogy to scamellum (: scabillum, to *scabnom scamnum). Meillet considers the word to be of unknown etymology; see Ernout-Meillet, Dict. Etym. s. v.
page 35 note 14. CGL 5.122.29.
page 35 note 15. Both pāgella and pāginula were probably coined by Cicero.
page 35 note 16. One is justified in positing a form with -en- (1) because the historical form asellus shows an e, and (2) because the similar e of the old diminutive bellus is guaranteed as phonetic by OLat. dvenos.
page 35 note 17. That fiscella was an old household word we infer from Isidore’s gloss, forma ubi casei exprimuntur.
page 35 note 18. Fēmella is very rare and impresses one as a coinage (of Catul. ?).
page 35 note 19. See note 1, supra.
page 36 note 20. For the equally possible analogical origin of columella, see note 6. To derive columella from *kelomen(e)lā necessitates postulation of syncope in the fourth syllable, which, in the judgment of some, would doubtless speak against the assumption of regular phonetic development for this diminutive.
page 36 note 21. Neut. pastillum Varro, after lībum (?); cf. Paul.-Fest. 251.13 M pastillum in sacris libi genus rotundi.
page 36 note 22. Even in antiquity pastillus was thought to be the diminutive of pānis; cf. Paul.-Fest. 222.18 M pastillus forma panis parvi, utique diminutivum est a pane. Sommer2 253 (and Krit. Erläut. 24) and Leumann 106 support this by deriving pānis from *pastni-s and pastillus from *pastni-lo-s. Walde2, however, would derive panis from *pā-nis (to pā-scor), and would take pastillus as diminutive to *pastum or pasta * dough ’ (< Gr. πάστη). See also Ernout-Meillet, Dict. Etym. s. v. pānis (<* spasms’).
page 36 note 23. Neut. pugillum CGL 2.164.30 (gender after Gr. δράγμα ‘handful’?); cf. 2.164.26 pugilla έπισφαίρια (influenced by pugil; cf. pugillō — pugilō).
page 36 note 24. By-form scabellum Varro; scamillus Vitr., scamillum Itala, and scamellum Prise, (also scamnellum Paul. Nolan.) are new formations on the historical primitive scamnum : scamnum (*scabnom) : scabillum (-ellum) = scamnum : scamillum (-ellum).
page 36 note 25. SiGiLLVM (VoLKANi) is found in Inscriptiones Fratrum Arvalium 357 (Marini).
page 36 note 26. A ‘ double * diminutive sigilliola -ōrum ‘ figurines ’ is found in Arnobius. A back-formation *sigula is to be assumed for sigla -ōrum ‘ signs of abbreviation ’, Cod. Just.
page 36 note 27. New formation tignulum Boeth.
page 36 note 28. Var. MSS reading batillum; by-form batillus Marc. Empir.
page 36 note 29. Vatillum, which means primarily ‘ shovel’, is probably to be referred, with Sommer2 232 (and Krit. Erläut. 86) and Leumann 143.b (PIE and PIt. -tn-), to vannus, inasmuch as the resemblance of a shovel to a winnowing fan is a close one. The divergence in gender is hardly a serious obstacle in the way of this assumption, since vannas was probably not originally feminine (its secondary diminutive vallus < *vann(e)los ‘ winnowing fan ’ is feminine, having been made after the primitive acquired its new gender); see Leumann 171.A.3; vatillum is neuter perhaps after ventilābrum of the same meaning. If vătillam is taken, with Petersson, IF 24.254-5, Walde2 s. v., and Muller, Altit. Wörterb. 522, as diminutive of a *
āt-elo- to vās < *
āts, the transition from the meaning ‘ vessel ’ to ‘ shovel’ proves difficult, and the short ă of vătillum must be explained as representing a different ablaut-grade from the ā of vās (: vătaxy vătius), or as a shortening of quantity in a pretonic syllable (cf. mōlēs mŏléstus and Leumann 77, where the assumption of shortening is disallowed; for pūsus pŭillus, see Sommer2 130). Meillet, Dict. Étym. s. v., would explain vatillum (ī ?) as a trade-name adapted from that of a manufacturer of goblets; cf. Mart. 14.96.1 (calicem Vatini) and 10.3.4 vătīnia ‘ drinking-cups (manufactured by Vatinius)’ (but cf. Friedländer’s note ad locum). Such a transference might have been possible in the case of articles such as goblets, but shovels are perhaps too inartistic a commodity to warrant their bearing a manufacturer’s name !
page 37 note 30. Only scabillum has by-forms in -ellum (see note 24), perhaps because in the meaning ‘ clapper (instrument composed of two boards moved by the foot) ’ this diminutive would naturally have been associated with tabella(e), and the suffix -ellum thus acquired would then have been extended also to cases where the word (or its by-forms) meant 4 bench ’; cf. Cagnat-Chapot, Manuel d’Archéologie Romaine 2.211, and see Daremberg-Saglio, Dictionnaire des Antiquités, figg. 5504 and 6142. For Republican authors the MSS tradition attests scabillum in Cato de agr. 10.4 (‘bench’) and Cicero Cael. 27.65 (‘clapper’), scabellum in Varro LL 5.168 (‘bench’); for the meaning ‘ clapper’, cf. also Suet. Cal. 54, Arnob. 2.73, and August, de musica 3.
page 37 note 31. See Leumann 106. PIt. ṇ is found only before Z; for criticism of the theory that the group -ṇl- became -enl- > -ell-, and that -ell- became -ill- primarily through the influence of -ll-, see § 15 supra, where also other unsatisfactory views have been treated. Attempts to explain the suffixal i of these diminutives as due to vowel- assimilation (*sigellum *tigellum becoming sigillum tigillum) or to analogies (pugillus after pugil ‘ boxer ’) must be judged unsuccessful; it is not at all likely that our six diminutives are analogical formations, since only one possesses the original ‘ phonetic’ form, viz. scabillum (: ‘ phonetic ’ scabellum).
page 37 note 32. This is not a serious difficulty, since the Latin products of the PIE syllabic liquids and nasal all differ from the products of
,
, and
arising in primitive Italic by samprasāraṇa : PIE
gave Latin -or-, -ol-, -en-, PIt.
,
,
gave Latin -er-, -il(l)- or -ul-, and -in(l)-; see Leumann 33, 35, 102.2, and 106.
page 37 note 33. In Osean and Umbrian Pit. ṇ apparently gave -en-; see von Planta, Gram. d.
osk.-umbr. Dialekte 1.323, and cf. Ose. patensíns perhaps < *patn(e)síns ‘ aperirenť (otherwise Muller, Altit. Wörterb. 323 s. v.
), and Umbr. Padellar ‘”Patellae ’ perh. < *Patn(o)lā- (cf. Osc. Patanaí and see von Planta, op. cit. 1.265-6, 2.30 and 260). The derivation oí Umbr. Afillis ‘ Amnius ’ (?) from *afni-lo- is highly uncertain (von Planta 1. 302). Mention should be made here of Sabellī < *Safn(o)loi, to Osc. Safinim ‘ Samnium ’. Sabellī perhaps shows secondary -lo- in a rare non-diminutive function; in any case the word has a dialectic e but a Latin b, not f (cf. Sabīnī < *Safnīnoi, Samnium < *Sabniom < *Safniom). Cf. Schulze, Eigennamen 481 note 8; Sommer2 7 and 83; and Leumann 172 .IX.H. 3 .a.u.![]()
page 38 note 34. See note 33 end.
page 38 note 35. See notes 24, 30, and 31 end.
page 38 note 36. For columella see List 2.3 and note 6 supra. Brugmann’s fallacious supposition, Gdr.2 2.1.367 footnote, that PIt.
became -in- only after g, was seemingly strengthened by columella and scabellum (pastillus and vatillum were not regarded by Brugmann as diminutives made on n-stems); for criticism see § 16 supra.
page 38 note 37. It is so explained by Leumann 106.
page 38 note 38. The primitive pānis is derived as follows : *pastnis > *pasnis » *paznis >pānis.
page 38 note 39. Popular Latin anguīla, which Sommer (Hdb.2 205) is inclined to believe the original form; anguilla would then show expressive gemination (see Lindsay, ALL 8.442).
page 38 note 40. For the suffix -īno/ā- and the generic adjectives and nouns formed by it, see Brugmann, Gdr.2 2.1.275 and 665; for the diminutive meaning apparent in some Latin derivatives in -īno/ā-, see von Planta, ALL 12.370-1. Several etymologies have been proposed for anguīlla, for which see Hofmann s. v. anguis. Although it is possible that anguīlla may be a ‘ relational ’ diminutive formed directly on anguis (see § 22 supra and note 4), I should propose rather that the word is a substantivized adjective (gender after vipera ?) which was formed on *anguīno- plus secondary -elo- in a ‘ generic ’ non-diminutive function; cf. arietīllum (sc. ālium) ‘ leek ’, which was formed on arietīnus ‘ belonging to a ram ’ in the same way. For the relation between ‘ generic ’ and diminutive meaning in the same formans, cf. OHG kindisk ‘ childlike, childish ’ with Gr. παιδίσκος ‘ little boy ’, Lat. mariscus (to mās, măris)
4 masculine * with Gr. άνθρωπίσκος ‘ little man ’, and Gr. άνθρωπικός ‘ of a man, human’ with Lat. homunciō (<*hemon-k-iō) ‘little man’; cf. Brugmann, Gdr.2 2.1.502-3, and Brugmann-Thumb, Griechische Grammatik4 (München 1913) 212. In Latin secondary -elo/ā- (-k-elo/ā-) was of course almost exclusively diminutive in function; but cf. masculus (-k-elo-) ‘ masculine ’, and see Friedrich, Demin. Bild. 70-1.
page 39 note 41. Attested only in the plural until Phaedrus (Ernout-Meillet s. v.). By-forms : armellum CGL 2.25.35, and armeīlae CGL 5.268.9. Armīllae was originally an adjective; perhaps spīrae or lāminae is to be supplied (cf. Cagnat-Chapot, Manuel d’Archéol. Rom. 2.358).
page 39 note 42. For ancient derivations see Paul.-Fest. 25.7 M (armillas ex auro, quas viri militares ab imperatoribus donati gerunt, dictas esse existimant quod antiqui humeros cum brachiis armos vocabant) and Priscian, GL 3.462.3 3 (from arma). The feminine gender of armīllae remains unexplained if the word is taken as a ‘ relational ’ diminutive made directly on armus (cf. Eng. armlet and Fr. bracelet 4 armlet ’ < OFr. bracelet ‘ little arm ’; see Meyer-Lübke, REW2 no. 1254). Armīllae was perhaps formed on *armīno- plus secondary -elo- with ‘ generic ’ function, and meant originally ‘ (circlet) belonging to, i. e. worn on the (upper) arm ’ (cf. bracchiāle), Friedrich, Demin. Bild. 60, suggests that armilla (fem. sing, with short ĭ) may be a translation of Gr.
‘ bracelet ’.
page 39 note 43. Cf. CGL 2.25.32 armilium σκε
ος.
page 39 note 44. Like armīllae, armīlium (sc. vās) may have been formed on *armīno- plus secondary -elo- in a non-diminutive function, and thus have meant originally ‘wine- vessel belonging to, i. e. carried on the shoulder’; one may suspect, however, that the extant Latin form has been reshaped from an unknown word in order to suggest armus (see Hofmann s. v. armilium). On the basis of the gloss in Paul.-Fest. 2.5 M armilium vas vinarium in sacris dictum (quod armo, id est humero, deportetur), this diminutive has been referred to Umbr. arsmor ‘ ritus ’; see Muller, Altit. Wörterb. 43 s. v. ăr[e]s-mos, and Ernout-Meillet s. v. armus.
page 39 note 45. Cf. Paul.-Fest. 20.19 M arbilla arvina, id est pinguedo corporis. New formation : arvīnula Vulg.
page 39 note 46. The noun bovilla (CGL 2.31.2 βοοστάσια) stands for bovilia (Götz, -illa trad.), pl. of bovīle (= būbīle) ‘ cattle-stall ’.
page 39 note 47. Both bovīnus and bovīllus are first attested in Livy, and then not until Late Latin; they were secondary formations on the stem bŏv- and did not maintain themselves against būbulus ‘of cattle, ox- ’.
page 39 note 48. Neut. catīllum Petr., to catīnum (gender after vās ?), fem. catīlla CGL 3.324.
70; cf. also Diomedes, GL 1.326.7 (antiqui) descendebant enim ad tertiam usque formanı, tam quam… catinus catinulus catellulus (read catillulus ?). Derivative : catīllo -ōnis ‘trencherman’, Lucii, and Paul.-Fest. New formation : catīnulus Varro ap. Charis.
page 40 note 49. The primitive is also spelled crumēna (< Gr.
(ί)α, -αία ‘ purse ’ ?; see Hofmann s. v. crumīna’).
page 40 note 50. Ovīnus is not attested until Seren. Sammon.
page 40 note 51. New formation : pulvīnulus ‘ bed of earth ’, Colum.
page 40 note 52. The masculine occurs in the proper name Rēgīllus; the feminine is found in Plaut. Epid. 223 (quid erat induta ? an regillam induculam an mendiculam ?), and as a technical term ([rēgīlla tūnica or vestis) in Varro ap. Non. 539.9 Mand Paul.-Fest. 286.33 M. The masculine form occurring in CGL 2.256.18 (regillus βασιλικός) is to be considered an extension of this originally feminine adjective, while rēgīllus of CGL 2.462.35 (regillus υιός βασιλέως) is a late diminutive to rēgulus -ī ‘princeling’ (cf. Livy 42.24.10 for rēgulus in the meaning ‘ king’s son ’); cf. also CGL 4.385.50 and 5.609.18 regillus regulus.
page 40 note 53. Rēgīna, an old ī-stem (< *rēgī-nā; cf. Skr. rājñī ‘ queen ’, mase, stem rājñ-), is probably to be assigned as the primitive, since the adjective was used primarily in connection with women’s clothing (see note 52). It is hardly likely that rēgīllus was made directly on regō (Ernout-Meillet s. v. regō; cf. only 8.5 miscellus, dubiously to misceō) or on rēgius (Non. 539.10 M), since *rēgiolus, not rēgillus, would then be expected, and since formations such as 8.6 novellus (to novus) and 9.13 putillus (to putus) were very rare.
page 40 note 54. Trin. 492 salillum animai… quom extemplo emisimus (satillum cod. A, scin- tillulam coni. Fleckeisen).
page 40 note 55. Suīnus is attested only in Varro RR 2.4.8 (-ina cod. e, -ilia edd.), Theod. Prise. 4.1 sevum suinum, and CGL 4.179.17 suinum porcinum.
page 40 note 56. Catellus (ē?) of Plaut. Curc. 691 may be a pun on catellus ‘(little) dog ’ (delicatum te hodie faciam, cum catello ut accubes, ferreo ego dicam), or perhaps the meaning ‘ chain ’ was momentarily extended to catellus from catulus, which, like canis, sometimes had this meaning (see Lucil. ap. Non. 36.26 M). New formation on catēna : catēnula Paul. Nolan.
page 41 note 57. Stilla, according to Meillet s. v., shows expressive gemination and is to be referred to Gr.
‘ bit, mite ’. This derivation is perhaps favored by the discrepancy in the meanings ‘ (viscous) drop * (stīlla) and ‘ icicle * (stīria); cf. Suet, frg. 319.27 Roth : gutta imbrium est, stilla olei vel aceti.
page 41 note 58. See notes 40, 41, 42 and 44.
page 41 note 59. Cf. the reverse combination in OHG -līn (kindilin) = NHG -lein (Kindlein) < -lo- + -īno-.
page 41 note 60. Both bovīnus and bovīllus were probably analogical formations (after suīnus suīllus), since būbulus was the usual word for ‘of cattle, ox-’.
page 41 note 61. See notes 47, 50 and 55.
page 42 note 62. Cf. Paul.-Fest. 286.33 M regillis tunicis, albis, et reticulis luteis utrisque <re>ctis, textis susum versum a stantibus, pridie nuptiarum diem virgines indutae cubitum ibant ominis causa; ut etiam in togis virilibus dandis observan solet. For the generalized meaning ‘ queenly, regal see Plaut. Epid. 223 quoted above in note 52.
page 43 note 1. Paul.-Fest. 24.17 M auxilia olla parvula. New formation on aulla (ōlla) : aullulla (ōllula) Plaut, or Varro.
page 43 note 2. See Hofmann s. v. aulla. The development of the primitive was : *aukslā > *auslā >*auzlā > aulla; for the writing -ll- see Sommer2 251 Anm.
page 43 note 3. Cf. the metathetic forms ascilla Veget. and ascella Itala (ascellus Ps. Diosc., astella Ps. Soran., arcella locus sub brachia CGL 5.169.8), in which x = -cs-) of axilla became -sc-. This change was assisted inasmuch as popular etymology referred axilla to ciliō -ere ‘ move ’ (for the authenticity of this verb, see Hofmann s. v.); cf. Isid. Orig. 11.1.65 (alas) quidam ascillas vocant quod ex his brachia cilluntur id est moventur. Cf. further the by-form auxilia : CGL 5 . 346.41 a[u]xillae alae minores; and the double formation ascellula : CGL 5.169.9 ascellulae pinnae vel squamae.
page 43 note 4. For the proper name Axilla see Schulze, Eigennamen 420; the word was very old as a cognomen of the gens Servilla (cf. C. Servilius Axilla, tribunas militum in 419 B. C).
page 43 note 5. The development of the primitive was : *akslā > *aslā > *aziā > āla with compensatory lengthening; see Sommer2 251.
page 44 note 6. By-forms : bacillus Isid. and bacellus Hieron.
page 44 note 7. By-form : imbēcillis (after dēbilis) Cic.
page 44 note 8. Baculum is to be derived with Hofmann s. v. from *bak-tlom (cf. Gr.
) > *bakklom > *baklom *bakolom (with anaptyctic ō). The simpler derivation of Baehrens, Sprach. Komment, z. vulgärl. App. Probi 17, from *bak-elom (with primary -elo-) is invalidated by imbēcillus, where -bēc- stands for -bĕcc- weakened from -băcc- : *en-bak-tl-elos » *
*imbakkillos > *imbeccillos > imbēcillus (-ē- Lucr. and Hor., -ĕ- Prudent.; cf. Sommer2 203-4). By the same token, *bak-tl-elom > *bakkḷlom was not simplified to *bakḷlom in the primitive Italic period, but became Latin *baccillom and then bācíllum (by the mamílla-law, which did not entail compensatory lengthening).
page 44 note 9. So Reichelt, KZ 46.336; for further literature on capillus see Hofmann s, v. OE hafola is from PIE *qap-elo- (beside PIE *qap-ut-), a stem which is not found in Italic except perhaps in capillus. According to an unpublished etymology of R. G. Kent, capillus is a compound and is to be derived from *kapo-pil-elo-s ( (small) head-hair ’ > *kappillos > capillus (by the mamílla-law); one might object that this would necessitate a double syncope and that a *kapo- (: kap-ut-) cannot be extracted from anceps (probably from *amb-kapot-s; Leumann 75 [i und 0 vor zwei Konsonanten]) or from prīnceps *prīzmo-kaps, to captō; Leumann 78.1. d). It is conceivable that capillus is a ‘ relational ’ diminutive formed on a supposed stem *cap- or that the word has been reshaped from an unknown form in order to suggest caput (cf. tranquilla, to -quiēs); such a popular association is apparent in Plaut. Epid. 623 ab ung<uic>ulo ad capillum summum.
page 44 note 10. By-forms : cavillum Paul.-Fest. (neut. after iūrgiumi), cavillus Apul. (mase, after iocus ?); cf. also Plaut. Truc. 686 cavillibus (comic for caulibus; cf. Truc. 683 caullator ‘ jester’), and CGL 2.98.34 cabillem διοασκαλίαν (contaminated).
page 44 note 11. Reference of cavilla to calvor was first proposed by Priscian, GL 2.431.21. ‘Kal
-lolā > *
(not *kalulolā, because of *kal
ōr) > *kalvillā > cavilla (with dissimilatory loss of l; see Sommer2 212). The suffix -lolā is perhaps not double diminutive -(e)lo/ā- but doubled primary (verbal) -(e)lo/ā- (for primary -lo/ā- without connecting vowel, cf. sella *sed-lā and iugulum * collar-bone ’, where u before l is to be considered anaptyctic in view of Gr.
). Assumption of a doubled primary -lo/ā- must remain uncertain because of the paucity of the evidence (besides cavilla, cf. favilla and scintilla); is it possible that the verbal abstracts in -ēla (querēla suādēla) show doubled primary -ĕlā- and that original *-ĕllā- became -ĕlā- in order to distinguish the verbal suffix from the ‘ double ’ diminutive suffix in -ĕllo/ā- ?? See, however, Leumann 172.IV.C on this debated suffix.
page 44 note 12. Cf. CIL 5.3143 favIlla (-ī- because of contamination with diminutives from stems in -īno- ?).
page 45 note 13. So Walde2 s. v. favilla, after Thurneysen, KZ 28.159. The PIE form perhaps shows, like *kal
-lolā cavilla, doubled primary -lo/ā-; early Latin *fovillā became *fovilla favilla (cf. *lov
re > lav
re and see Sommer2 109-10). Favilla never acquired a *favillula, as cavilla did cavillala (Plaut.).
page 45 note 14. To be referred perhaps to Popular Latin magulum ‘ mouth ’ (Schol. Juv. 2.16); but cf. Sommer2 231 (māla <*smak-s-lā-, to Lith. smak-rà ‘chin’). Maxilla is dubiously derived by Thurneysen, IF 21.177, from *mand-slelā (to mandō ‘ chew’) becoming *mansillā and then maxilla by analogy to axilla; for criticism see Sommer, Krit. Erläut. 84.
page 45 note 15. Var. spelling paus illus (analogy to pusillus ? see Sommer2 248). Double formation : pauxillulus Plaut, (pausillulam Plaut. Stich. 163 cod. A). Because of the dissimilarity of paul(l)us and its diminutive pauxillus, the former early acquired the new formation paul(l)ulus (Plaut.).
page 45 note 16. See Walde2 s. v. paullus. The development of the primitive was : *pauk-slo-s *pauslos > *pauzlos > paul(l)us; for the original writing with -ll- see Sommer2 251 Anm.
page 45 note 17. To pangō (= *pag- with nasal infix); see Walde2 s. v. pālus.
page 45 note 18. Mase, pistillus is more frequent (gender after calx ‘ mortar ’ ?); pistillum Plaut. Aul. 95 (-ellum cod. B) and Non. 221.7 M.
page 45 note 19. To pĭstus, perf. pass. part, of pīnsō ‘ pound ’ (-n- is a nasal infix belonging only to the present system; short ĭ of pĭstas is attested by Prov. pestar ‘ pĭstāre * and OFr. pestar ‘ pĭstor ’ : see Meyer-Lübke, REW2 nos. 6536 and 6537). For the derivation of pīlum from *pist- 1o-, see Leumann 144.e. α end; cf. Sommer2 259 (and Krit. Erläut. 23), who would derive the word from *pīns-tlom (to pīnsō). Sommer’s form would have become *pīsculum, and pīstillum would then necessarily show -ī-. An objection to Leumann’s form is that it is made on the participial rather than the present stem; *pist-lom did not become *pisculum, because -t- was not suffixal, i. e., was maintained by association with *pistos.
page 45 note 20. Pōclum is read by Lindsay for Plaut. Asin. 771 (poda codd., d misread for cl in the archetype), Curc. 359, and Pers. 775a. Pōculum (with anaptyctic u) is from *pōklom < PIE *pō-tlom (cf. Skr. pā-tram ‘ cup ‘).
page 45 note 21. Pŭsillus is not from pūsus ‘little boy’ (pūsillus would be expected!). The group -tsl- had a different development from that of -ksl- (cf. āla with compensatory lengthening from *akslā) : *pŭt-slo-s (to Skr.
‘ son ’) > *pusslos > with assimilation pŭllas (-ŭ- attested by Romance derivatives : see Meyer-Lübke, REW2 no. 6828); the diminutive developed differently : *pŭt-sl-elo-s » *pŭssḷlos > *pŭssillos > pŭsillus (assimilation was precluded at the second stage because of the vocalic nature of
). See Sommer2 251-2 (and Krit. Erläut. 86-7), and Leumann 144.e.α end; for other members of this group see Walde2 s. vv. pūsus and pūtus.
page 46 note 22. Original gender uncertain; Cato Agr. 133 used the diminutive in the masculine (eum quasillum… impleto), and Agr. 11 the primitive in the neuter (quala satiiria vel álveos xl); mase, quasillus was perhaps influenced by calathus (cf. Paul.-Fest. 47.6 M), and neut. quasillum (Cic.) by canistrum or cōlum ‘ strainer ’.
page 46 note 23. Quallus has MS authority in Verg. Georg. 2.241 (qualus edd.), but qualus or qualum is the usual spelling found in MSS (= quā-; the vowel-length, metrically assured, shows original gemination of l).
page 46 note 24. Apparently to quatto; see Leumann 144.e.α. According to Sommer2 252 Anm. and Walde2 s. v. quālum the stem is *quas-s-lo-. In either instance, the development of quăsillus and quăllus was parallel to that of pŭsillus and pŭllus (see note 21).
page 46 note 25. Only with hesitation is scintilla to be derived from *sqi-n-t-lolā, to root *(s)qāit- ‘ bright, clear ’ in ON held ‘ unclouded sky ’ (cf. NHG heiter ‘ clear ’) and probably in Lat. caelum (, *cait-[s]lom; see Hofmann s. v.). The suffix -lolā perhaps represents doubled primary -lo/ā- (cf. cavilla and favilla above), although scintilla was doubtless felt to be a diminutive (cf. the ‘ double ’ formation scintillala Cic.). Niedermann, IF 26.58-9, sought to show that both scintilla and Gr.
were borrowed from a non-Indo-European Mediterranean language; in this case scintilla would have acquired its ending by analogy perhaps to favilla.
page 46 note 26. A denominative adverb (to singulus or *singillo/ā-, not *singillō -āre) like paulātim (to paulum); see Leumann 215 .c.
page 46 note 27. According to Leumann 171.B.2.b, singulus is from *sevglos dissimilated from *wn-gnos ‘ born one (apiece) ’ (= Goth, ainakls ‘ alone ’ < PIE *oino-gnos). Accordingly, *singillo- may be derived from *sevgl-elo- (adjectival -elo-) » *
- > *sin- gillo-. For another derivation of singulus see Walde2 s. v.
page 46 note 28. Taxillus may best be reckoned as an analogical formation after maxilla or paxillusi since the derivation of tālus is highly uncertain (see Walde2 s. v.; hardly with Muller, Altit. Wörterb. 469 s. v. ta[g]zlos, to tango, with an original meaning ‘ *hollow of the knee ’).
page 46 note 29. Var. MSS reading tōsillae; cf. mōstellum (List 1.26).
page 46 note 30. To tēnsus (or *tēnsio = tendō); see Sommer2 251. Tolls is from *tons-lēs > * tozlēs > tōlēs; tōnsillae is from *tons-li-lai > *tōns
lai > tō(n)sillae (-ns- was maintained at the second stage because of the vocalic nature of
). Cf. Paul.-Fest. 357-9 M.
page 47 note 31. Vixillum is attested by some MSS for Plaut. Trin. 888 (cf. also CIL 6.1377), iuxillum ibid. cod. B; cf. also CIL 3.79 vexillvs (mase, after pannus ‘ cloth ’ ?).
page 47 note 32. See Walde2 s. v. vēlum (to root *
eg- ‘ weave’).
page 47 note 33. See note 5.
page 47 note 34. Pullulus ‘ chick ‘ (Apul.) was a new formation on pullas ‘ chicken ’.
page 47 note 35. Cf. a different development of PIt.
in facillimus < *fakellemos <*fak
semos < *fakli-semos (-il- in facillimus because of facilis ? but see Leumann 210.b and 102), and in facultās < *fakoltās < *fakeltās < *fakiḷtās < *fakli-tāt-s (to facilis < *fak-li-; or facultas was formed upon facul ‘ easily ’ < *
<* faklı). The evidence of Oscan and Umbrian is too scanty to enable one to conclude that the product of ḷ was el before l in these dialects; cf. Osc. Kupelternum < *Kuploterno- (?) and the Osean town-name Abelia *< *ablonā ‘ *apple-raising ’ (so Hofmann s. v.). In Ose. Mítl ‘ *Mitulus ’ and Fíml ‘ *Fimulus ’ the writing l may represent el (not ḷ), in view of Osc. famel ‘ famulus’ and Umbr. katel ‘ catulus’. See von Planta, Gram. d. osk.-umbr. Dialekte 1. 323-4.
page 48 note 36. See notes 9, 11, 13, and 25.
page 48 note 37. Axilla is very rare until Late Latin, and both its meanings were represented in the primitive āla (for the meaning ‘ arm-pit * in āla, cf. its derivative ālipilus).
page 48 note 38. See notes 14 and 28.
page 48 note 39. A primary suffix is one which forms derivatives of radicals or of verbal stems (e. g. trem-ulus, teg-ulum), a secondary is one which forms derivatives of nominal stems (e. g. flōs-culus, paul-ulum, and mās-culus). In Latin secondary -elo/ā- was almost exclusively diminutive in function.
page 48 note 40. Fem, agnella Ennod., to agna.
page 48 note 41. *Agn-elos would phonetically have become *agillus (see List 3). Agnulus was a Late Latin formation on agnus; agnellus (Plaut. Asin. 667 and Late Latin; cf. Fr. agneau) owed its ending to vitellus. For avillus(to ovis rather than to agnus) see List 9.2.
page 48 note 42. Var. MSS reading annellus.
page 48 note 43. So Hofmann s. v. angulus, in view of Skr. angu-liḥ ‘ finger’ and Gr.
‘ bent ’. Umbr. anglome ‘ ad angulum ’ is a syncopated form and does not justify positing a stem * anglo- for Latin angulus, as is proposed by von Planta, Gram. d. osk.- umbr. Dialekte 1.554 and 2.25, and Leumann 151. d. The suffix is here secondary; cf. ancus ‘ qui aduncum brachium habet ’ (Paul.-Fest. 19.15 M), ‘ elbow ’ in Romance languages (Meyer-Lübke, REW2 no. 446), and see Hofmann s. v.
page 48 note 44. Area was also used in this technical sense; see W. van der Goes, Rei Agrariae Auctores (Amsterdam 1674) 119.
page 49 note 45. Amella, which according to Varro LL 8.79 was disused, is next attested in Apicius; cf. AVCVLA CIL 6.33706, whence the back-formation anca ‘ goose ‘ (Avian.). Mase, aucellus CGL 2.25.42. New formation : avicella Mart. Brae., to avis or *avica > auca.
page 49 note 46. The stem is strictly -k-elā-; for the distribution of -ulo/ā- and -culo/ā- in historical Latin, see § 2 note 2, and Priscian, GL 2.102.20.
page 49 note 47. Paul.-Fest. 35.9 M blandicella verba blanda per diminutionem sunt dicta (= CGL 5.520.25; both glosses refer to a lost passage in Plautus ?). Blandicella is an early example of -kello/ā- attaching to an o-stem and may have been patterned after molli-cellus (or vice versa ?).
page 49 note 48. Fem, catella Petron., to catula.
page 49 note 49. Cf. Umbr. katel (< *katel(o)s; the gen. katles is syncopated) and MHG batele ‘ goat ’; these words, together with catulus, probably have non-diminutive -elo-. Attempts to relate catulus to canis (Varro LL 9.74) or to cattus ‘ (wild) cat’ (first found in Pallad.) have been unsuccessful; see Hofmann s. v. catulus.
page 49 note 50. For the emotional coloring of this and like diminutives in Plautus, see F. Conrad, Glotta 19.136 and 140-1 (1931).
page 49 note 51. Alfenus ap. Digest. 8.5.17.1. Otherwise the attestations are very late : cf. cocumella Papy. Marini 80.2.11 (564 A. D.), with -0- because of analogy to coquō ‘ boil water ‘; and cucumellutn Optat. (neut. perhaps after lābellum ‘ bathing-tub ’; for cucuma with this meaning see Mart. 10.79.4). Cucumula is first attested in Petronius.
page 49 note 52. Perhaps the reading of Plaut. Most. 895 : culcitullam codd., -ulam Leo and Lindsay, -ellam some other edd.
page 49 note 53. For the suffix -blo/ā- < PIE -dhlo/ā-, see Leumann 172. V.C.
page 49 note 54. Var. MSS reading elvella. *Heluella was originally perhaps a neut. pl. diminutive made on a supposed stem *helu- (to helus -eris = [h]olus ‘ vegetables, cabbage ’); cf. Paul.-Fest. 103.7 M helvela (-elia Lind.) olera minuta. When helus passed out of currency, *heluella was referred to the color-adjectives helvus and helv(e)olus and became helvella (if understood as a fem. sing. adj. it might have been associated with brassica ‘ cabbage ‘); for the phonetic change of -lu- to -lv-, ct. mīluos > mīlvus.
page 45 note 55. Var. MSS reading matēlla (contamination with muntēle ‘ napkin ’ ? or influence of cubīle, since matēlla meant also ‘ chamber-pot ’). Derivative : matelliō -ōnis masc.4 pot ’; see Leumann 176.II. A. 2.
page 45 note 56. No satisfactory etymology has yet been proposed for matula : see Walde2 s. v. and Muller, Altit. Wōrterb. 258 s. v. măt(ĕ)lā.
page 45 note 57. Strictly -k-elo-; see note 47.
page 50 note 58. To root *oq
- ‘ see ’; see Walde2 s. v. oculus. Nonce-formation : ocellulus Diom.
page 50 note 59. So Leumann 81.1 in view of (pilumnoe) poploe Fest, and poplicus ‘ publicus’; otherwise Walde2 s. v. populus.
page 50 note 60. Cf. Lith. parszēlis 4 porculus ’ and OHG farheli (= NHG Ferkel). Porcellus has the late derivative porcelliō -ōnis mase. ‘ sowbug, woodlouse ’; see Walde2 s. v.
page 50 note 61. Cf. the proper name Rūsticilla. In proper names the suffix -illo/ā- was far commoner than -ello/ā-; cf. Drūsilla, Flaccilla, etc.
page 50 note 62. Cf. Livy 25.3.16 sitella lata est ut sortirentur, and see Weissenborn’s note ad loc.
page 50 note 63. See Walde2 s. v. sīnum ‘ drinking-bowl’. Muller, Altit. Wörterb. 429 s. v. sīno-s, suggests that -tula in situla and matula (op. cit. 258) represents PIE -ilā- (but -tlā- regularly gave -culā- !).
page 50 note 64. See Walde2 s. v. tabula.
page 50 note 65. Cf. vitellus ‘ yolk of an egg ‘ Cic., -um Varro (cf. CGL 3.579.33 vitellus id est meiolus [= mediolus] ovi). Does the explanation of this meaning of vitellus lie in a notion of folklore, or was vitellus ‘ calf’ (-ĭ- attested metrically by Hor. Sat. 2.4.14 for the meaning ‘ yolk ’) taken for a diminutive of vīta and given the new meaning ‘ *bit of life > yolk’ ? The close association by countrymen of vitellus ‘ calf’ and ovis ‘ sheep’, together with the latter’s chance similarity to ōvum ‘ egg ‘, may have been the starting-point of this curious transference of meaning.
page 50 note 66. Cf. Ose. Viteliú ‘ Italia ’; Umbr. vitluf ‘ vitulōs ’ is syncopated. Vitulus may be a derivative of vetus ‘ old ’ (= Gr. (F)έτος ‘ year’; see Skutsch, ALL 15.35) plus adjectival -elo- and may have meant originally ‘ *yearling ’; see Walde2 s. vv. vetus and vitulus for parallels. A difficulty is that *viterulus would be expected, since vetus was an 5-stem; but cf. the Latin formation vetulus ‘ old(ish) to vetus -eris. Also the i remains unexplained (Varro LL 5.96 referred vitulus to Gr.
‘bull’ [Hesych.]).
page 51 note 67. The distinction between the diminutive endings -ulojā-, -ello/ā- (-illo/ā-), and -ellulo/ā- (-illulo/ā-) was one of emotional nuance rather than one of progressively diminutive meaning (‘ small … ‘ very small …’, ‘ tiny …’); see note 50.
page 51 note 68. It is true of course that diminutives were amplified or strengthened in different ways in different languages; cf. Gr. -ίσκιον, a combination of dimin. -ισκο- and dimin. -ιον, OE -kīn < dimin. -ko- + adjectival -īno-, and OHG -līn (= NHG -lein) < dimin. -(e)lo- + adjectival -īno-, See Cooper, Word Formation in the Roman Sermo Plebeius 168, for the cumulatory diminutive suffixes and reasons for their productivity; cf. also F. Conrad, Glotta 20.83.
page 52 note 69. Angellus is attested for the Republican period only in Lucr. 2.428.
page 52 note 70. Cf. Plaut. Asin. 667 and 691.
page 52 note 71. Anaptyxis in the suffix -bulo/ā- is older than any extant tradition; see Leumann 81. 1, and cf. CIL I2.581 tabolam. Tableis (CIL I2.585) and Umbr. tafle ‘ in tabula ’ are syncopated forms.
page 52 note 72. Popellus is found only in Hor. Ep. 1.7.65 and Pers. 4.15.
page 52 note 73. Apparently found only in Plaut. Capt. 137 foris aliquantillum (sc. frustum) etiam quod gusto id beat.
page 52 note 74. To amb(i) and colō (<*
- ‘ turn, move oneself’); anculus -a = Gr.
‘ servant ’; see Hofmann s. vv. anculus and colō. Juret, Dominance et résistance 158, would derive ancilla from *anklolā, to the syncopated forms anclō ‘ ministro ’ and anclābris (sc. mēnsa) ‘ sacrificial table ’. In my opinion aneilla has an analogical i; *am-
-elā would phonetically have become *anculla (cf. ampulla < *ampor-elā). With the obsolescence of ancula, aneilla was no longer felt to be a diminutive; ancillula is found as early as Plautus.
page 52 note 75. By-forms : anucella Fronto, to anucula CGL 3.439.50, and anilla Macrob. de differ, verb. 692, to anula; cf. CIL 8.7694.7 Anvlla.
page 52 note 76. Or -lo- (?) in view of Gr.
‘ enclosure ’.
page 53 note 77. By-forms : cōdicillum Cod. Just. (neut. after diploma ?) and cōdicellus Sen. (-e- after tabella ?). Cōdiculus (cau-) is not attested until Late Latin and may be a backformation.
page 53 note 78. Crepitācilhim has been included in this list rather than in List 5 on the ground that it was probably formed analogically on the historical primitive in -culum.
page 53 note 79. Frūstillum is first found in Arnobius (cf. frūstellum Marc. Emp.; -e- after of ella ?), and frūstulum first in Apuleius.
page 53 note 80. Derivative : furcillēs ‘ gallows ‘ Paul.-Fest. (3d declen, after crux crucēs ?).
page 53 note 81. Fem. haedilla Itala. Haedulus is first found in Juvenal.
page 53 note 82. Ōscillum is attested principally as a technical term from Roman religion; see Verg. Georg. 2.389 (cf. Serv. ad loc. and Macrob. S. 1.7) and Paul.-Fest. 194.8 M. Friedrich, Demin. Bild. 59, suggests that the word was a translation of Gr.
‘ small mask’.
page 53 note 83. To root *pap- ‘ swell ‘; see Walde* s. v. papula.
page 53 note 84. Neut. pēnicillum Quint, (gender after other neuter nouns denoting instruments); cf. CGL 3.207.50 penicellum σπóγγο<ς> (cf. Col. 12.18); see also 18.57.
page 53 note 85. Nowhere attested, and expressly disallowed by Varro LL 9.74 : cum sit anus cadus simile et sit ab anu anicula anicilla, a cado duo reliqua quod non sint propagata, sic non dicatur a piscina piscinula piscinilla.
page 53 note 86. Pūpilla was used in the meaning ‘ pupil ’ by Lucretius and Pliny sen., pūpula in the same meaning by Cicero, Varro, Catullus, Horace, and Ovid. Pūpilla with the meaning ‘ orphan ’ was perhaps originally merely the feminine counterpart of pupillus ‘ orphan ’, and later acquired the meaning ‘ pupil’ from pūpula (originally a ‘ relational ’ diminutive translating Gr.
‘ pupil ’).
page 54 note 87. Not attested until Late Latin (CGL 2.532.18; a back-formation ?).
page 54 note 88. Cf. Paul.-Fest. 332.2 M scrutillus venter suillus condito farre expletus. Scrūtulus is a late word (see CGL 5.611.32, where Götz reads scrutillus).
page 54 note 89. Cf. also scrūta -ōrum ‘ rubbish ’, and see Walde2 s. v. scrautum.
page 54 note 90. See note 76 supra; for crepitācillum (< *krepitā-kl-elo- ??) see note 78.
page 54 note 91. Double syncope may of course have taken place in the development of some words; see Götze, IF 41.84, who would derive diurnus thus : *di
esinos > *diverinos > *divernos > diurnus (equally possible is analogical formation after nocturnus, to noctū ‘ by night ’; cf. diū ‘ by day ’ and see Hofmann s. v. 1. diū). Derivation of -illo/ā- in these words from -lelo/ā- is hardly feasible, since the diminutive suffix was regularly -elo/ā- elelo/ā- if geminated).
page 55 note 92. Cf. perhaps Plaut. Trin. 799 servos ancillas amove, and Cic. pro Mil. 55 ducebat ancillarum greges.
page 55 note 93. It is hardly likely that cōdidllus was influenced by pugillārēs ‘ writing tablets ’, since the latter was probably first used in Imperial times.
page 55 note 94. Influence of ōscillum (‘ mask [suspended in the open air] ’) by a verb cilio ‘ move ’, though assumed by the ancients (cf. Fest. 194.8 M), is very dubious, since the existence of the latter has been called in question (see Hofmann s. v.).
page 55 note 95. See Juret, Dominance et résistance 154, for a similar explanation of pēnicillus. It is not likely that an original e in these words was ‘ palatalized ’ to i by the preceding c, inasmuch as we may offset ancilla, anicilla, bacillum, cesticillus, cōdicillus, cōnspicillum, crepitācillum, fureilla, opicillum, ōricillā, ōscillum, pēnicillus, pōcillum, and specillum by as many Republican diminutives with -ce- : arcella, aucella, blandieella, cancellī, fiscella, locellus, lucellum, macellus, miscellus, mollicellus, ocellus, por cellus (-a), pulc(h)ellus, rūsticellus, sacellum.
page 55 note 96. Cf. Plaut. Pseud. 179 ubi isti sunt quibu’ vos oculi estis, Quibu’ vitae, quibu’ deliciae estis, quibu’ savia, mammia (mamilla cod. P), mellillae ? See also note 61 supra. For the frequency of -illa in names of parts of the body, see Meillet, MSL 15.367.
page 55 note 97. See notes 77, 79, 81, 87, and 88 supra.
page 56 note 98. See Sommer2 109. Covella, being a ritualistic term, may have been conservative of the older vocalism. Juno Kalendaris was associated with the course of the moon and was invoked at the time of the new moon with the name Covella, which perhaps referred to the 4 hollow’ disc visible at that period; see also Varro LL 6.27 and G. Wissowa, Religion und Kultus der Römer2 (München 1912) 187.
page 56 note 99. Paul.-Fest. 105.12 M irnela vasis genus in sacris (irnela Lind.). In Cic. Parod. 1.2.11 the correct reading is urnulas.
page 56 note 100. Var. MSS reading manteliini (contamination with mantēle ‘ napkin ’ : Walde2 s. v.; -ellum in Romance : Meyer-Lübke no. 5326). Mantuni is attested in Isidore and in Romance (Meyer-Lübke no. 5328).
page 56 note 101. By-form miscillus Mart. Cap. Miscellus (with a passive meaning !) was referred by the Romans to misceō, as is evident from ludi miscelli Suet. Calig. 20 and from the derivatives miscellāneus Juv. 11.20 and miscelliōnēs (Paul.-Fest. 123.7 M m. appellantur qui non certae sunt sententiae sed variorum mixtorumque iudiciorum sunt). On the basis of technical passages in Cato, however, Leumann (Glotta 11.190) would take miscellus as a diminutive to minusculus ‘ rather small ’ : *minus- céllus > *minscéllus > mīscellus (cf. mīsterium = ministerium). He justifies this derivation by the fact that vintners pick the smallest grapes growing on poor vines for the vintage; thus ūvae mīscellae ‘ *poor grapes ’ came to mean ‘ grapes of mixed color ’ and the meaning ‘ mixed .. . ’ became completely generalized through the natural association of miscellus with misceō.
page 56 note 102. Used esp. of flora and fauna; cf. Cic. de fin. 5.14.39 arbor et novella et vetula, and novella -ae (sc. vītis) ‘ newly planted vine’ Coripp. Cf. perhaps also Ose. Núvellum ‘Novellum ’ (see von Planta, Gram. d. osk.-umbr. Dialekte 2.24; otherwise Gōtze, IF 41.84, and Muller, Altit. Wörterb. 294 s. v. nö
os). Juret, Dominance et résistance 154, proposes to derive novellus from *novero-lo-, to nover-ca (but see Walde2 s. v.). For novellus in Romance, see Meyer-Lübke no. 5967.
page 56 note 103. Var. MSS reading vuls-; volsellae occurs in Plain., Cels., and Mart., volsillae in Varro (i because of analogy to other nouns in -illo jā- denoting instruments; see § 34 supra).
page 57 note 104. Cf. Paul.-Fest. 20.12 M arillator codio qui etiam cocio appellatur, and CGL 5.7.38 arilla<tore> coactione (codione ?) panniculario; see Löwe, Prodr. 285, and Hofmann s. vv. arillātor, arrobō, and cōciō. Arillātor presupposes an *arilla < * arrilla, to arra, which was a short form of arrabō -ōnis masc. (Plaut.).
page 57 note 105. *Ovillus > avillus (cf. *cov
re > cav
re, and see Sommer2 109); since avillus was purely a Latin formation it can hardly be derived with Fröhde, BB 1.327-8, from PIE *ag
(h)n-elos, to agnus. And yet from ovi-s one would expect ovicula *ovicella.
page 57 note 106. Cf. specula ‘ watch-tower ’. Cōnspicillum is very rare and varies between verbal and diminutive meaning : for the former cf. Plaut. Cist. 90 dum redeo domum, Conspicillo consecutust clanculum me usque ad fores, and for the latter cf. Non. 84.5 M conspicillum unde conspicere possis. Plautus Medico : in conspicillo (-ilio codd.) adservabam, pallium observabam.
page 57 note 107. The original (unattested) meaning of fracillī may have been ‘ oil-lees ’ after that of the primitive fracēs. Since fracēs = floccēs ‘ lees (of wine) ’ and floccēs was confused with fioccī ‘tufts of wool ’, the latter meaning apparently passed into fracillī; cf. Paul.-Fest. 90.10 M fratilli villi sordidi in tapetis. Furthermore, since fraces was associated with fraceō ‘ putrefy, decay a corresponding meaning passed into the primitive and its diminutive; cf. CGL 5.22.11 fracebunt sordebunt, displicebunt, dictum a fracibus qui sunt stillicidia sterquilinii with CGL 5.22.23 fratellis sordium glomusculis (flat- cod. R; fracellis coni. Götz), 5.599.20 fritillum stillicidium stercoris in sterquilinium, and 5.653.16 fretillum fimum purgo. The gloss of Paulus quoted above may therefore represent an attempt to relate three disparate meanings or notions : villī represents floccī (: floccēs = fracēs), sordidi represents the notion in fraceō (: fracēs), and in tapētīs is doubtless corrupted from in trapētō ‘ in an oil-press ‘ in the original gloss (which perhaps read fracilli fraces in trapeto contusi). See also DuCange, s. vv. fracelli and fractellum. For the writing t for c before the palatal vowels e and i, cf. conditiō for condiciō and see Leumann 111 .d; for the variations in the initial syllable (fra-, fla-, fre-, fri-) cf. clītellae List 1.9 and its note.
page 57 note 108. Var. forms : fringuilla (Meyer-Lübke no. 3516), fringillus Varro, and frigilla (cf. frigō -ere ‘ squeak ’ and frigulō -āre ‘ scream’ [of the daw]).
page 58 note 109. Not from mel(l)ina ‘ mead ’, since *mellella would then be expected (see List 2). Melliila was used like a proper name and so is comparable to Septimillus (to Septimius : Catul. 45.1 and 13).
page 58 note 110. Cf. Fest. 139.5 M moracias nuces Titinius duras esse ait, unde fit deminutive moracillum (hardly to mora : see Walde2 s. v.).
page 58 note 111. Cf. Paul.-Fest. 158.17 M moscillis Cato pro parvis (pravis coni. Jordan) moribus dixit, and CGL 5.604.51 mosillus parvus mos. Keller, ALL 4.139 (and Lat. Etym. 73), would read the passage in Paulus as muscillis Cato pro parvis muribus dixit (but see Birt, Rh. Mus. 51.263). Muscillis is favored by the fact that mūsculus usually meant ‘muscle ‘ and almost never ‘ little mouse ’; in my opinion, however, moscillis should be retained, since the plural number plainly indicates that the word was made with reference to mōrēs ‘ (bad) character’.
page 58 note 112. Varro ap. Non. 83.25 M in uxorculae opicillum (gender after auxilium); the nearest morphological analogue of opicillum is perhaps ōri-cilla (to auri-s).
page 58 note 113. Pŭtus apparently occurs only in Verg. Catalep. 7.2 dispeream nisi me perdidit iste putus (pothus potus codd., putus Scaliger, πόθος Spiro). The (rare) formation pŭlillus was made on pŭtus because pŭsillus was felt to be the diminutive of pŭsus ‘ boy ’; see Walde2 s. vv. pŭsus and pŭtus and List 5.12, Meyer-Lübke no. 6890.
page 58 note 114. Fest. 274.8 M ravillae (raviliae Lind.) a ravis oculis quemadmodum a caesiis caesullae. Rāvillae may be an etymologizing of the proper name Rāvilla (see Hofmann s. v. caesius); otherwise why the fem. pl. ? Cf. Schulze, Eigennamen 219 (Ravillius, Ravelio<s>).
page 58 note 115. Hardly to suāvium ‘ kiss ’ (= sāv-; dimin. s[u]āviolum Catul.). Sāvillum is dissimilated from suāvillum (found in some MSS); gender after condīmentum or cāseum (neut. in Cato Agr. 76.3), or cf. the neut. substantive bellāria -drum ‘ dessert to bellus.
page 58 note 116. Not to speculum ‘ mirror ’; morphologically cf. cōnspicillum, to cōnspiciō.
page 58 note 117. Cf. Varro LL 7.65 strittabillas a strettillando; strinare ab 3 qui sistit aegre (cf. Meyer-Lübke no. 8318 strĭttāre 4 langsam einhergehen ’), and Ps. Plaut, ap. Gell. 3.3.6 scrattae, scrupedae (‘ hobbling ’), strittivillae sordidae. For this obscure word’s association with meretricious women and its possible corruption into strictivillae (to stringō and villus), see Keller, Volksetvm. 178.
page 59 note 118. Cf. Plaut. Cist. 407 diobolares, schoeniculae, miraculae, Cum extertis talis, cuw todillis crusculis (see Lindsay’s critical note ad loc.), Paul.-Fest. 353.2 M todi genus avium parvarum. Plautus: Cum extortis talis, todillis crusculis, and CGL 5.624.39 todillus gracilis. Todillus (ī ?) may hardly be referred to todinus (ī ?), which is attested only by Priscian. See 10.7 cro[co]tillus.
page 59 note 119. See notes 100 and 102.
page 59 note 120. See Wölfflin, ALL 12.305 (Analogiebildungen auf -ellus, -ella, -ellum).
page 61 note 1. See Thes. Ling. Lat. s. v. and cf. Paul.-Fest. 7.10 M ab alterno altellus, CGL 2.565.51 altellus alius diminutivum, and CGL4.206.10 altellus terrae nutri tus (contamination of alō and tellūs).
page 61 note 2. See Hofmann s. v.
page 61 note 3. See Hofmann and Ernout-Meillet s. v.
page 61 note 4. The word may have entered Latin as casmillus (with expressive gemination) and then have become *cammíllus > Camillas; see Hofmann s. v. and Wissowa, Relig. u. Kult. d. Römer2 496.
page 61 note 5. See Hofmann s. vv. cella and celo. The diminutive cellula occurs as early as Terence.
page 61 note 6. Crocotillus is not connected with crocōtula -ae ‘saffron robe ’ (Plaut. Epid. 231). Cf. Paul.-Fest. 52.20 M crocotillum valde exile. Plautus : … crocotillis crusculis with 353.2 M… cum todillis crusculis (Fest. 301.27 M sodellis); cf. also Fest. 300.12 M succrotilla vox tenuis et alla. Titinius : feminina fabula succrotilla vocula. See todillus (9.18) and its note. In the phrase cum todillis crusculis, todillīs was perh. made alliterative with crūsculīs and altered to *crotillīs. The compound suc-crotilla may then have been modeled upon sub-tīlis (whence also the s- of sodellīs ?). Was *crotillus later contaminated with crocodīlus *-illus ‘crocodile ’ and enlarged (metri gratia also ?) to crocotillus ?
page 61 note 7. Cic. ad Att. 13.51.1 (reading dubious).
page 62 note 8. Paul.-Fest. 105.14 M ipsullices bracteae in virilem muliebremque speciem expressae (ipsilles coni. Bugge, Jahrb. klass, philol. 105.99).
page 62 note 9. See Walde2 s. v.
page 62 note 10. See Niedermann, Ĕ und Ĭ 65-6, who would derive (unconvincingly) mellum from nom. or acc. *ménlom, and the form with i from an oblique case : gen. *mṇlei > *menleī > *mellī >millī (e > i before ll; see op. cit. 69).
page 62 note 11. Cf. Varro LL 5.76 motacilla quod semper movet caudam; see Hofmann s. v. ciliō, and Keller, Volksetym. 52.
page 62 note 12. See Walde2 and Ernout-Meillet s. v.; l was perhaps geminated after the incisive Gr. υ, a sound sometimes represented by i in Latin; cf. brabilla (beside brabyla)-:Gr. βράβυλος fem., serpillum = Gr.
ρπυλλον, and mītulus — Gr.
.
page 62 note 13. See Walde2 s. v. mustēla and Keller, Volksetym. 46 (popularly taken as musdela — quae mures delet ?).
page 62 note 14. See Walde2 s. v. (etym. uncertain).
page 62 note 15. See Walde 2 s. v. (s- after serpō).
page 62 note 16. See Walde2 s. v.
page 63 note 17. See Ernout-Meillet s. v. (‘composé expressif qui évoque quiēs’); but see also Walde2 s. v. and Muller, Altit. Wōrterb. 493.
page 63 note 18. This word was perhaps felt to be a diminutive as early as Plautus (cf. Pseud. 109 Lind. : seis tu quidem hercle, mea si commovi sacra, Quo pacto et quantas soleam turbellas dare), and certainly by the time of Apuleius, who took the word as a diminutive of turba ‘ crowd’. See Leumann 172.IV.C and List 5.5, note 11.
page 63 note 19. Vīlla is from *
eik-slā (cf. vīcus < *
eikos < *
oik-os); see Sommer2 251.
page 63 note 20. See Walde2 s. v. and also Niedermann, Ĕ und Ĭ 66-7.
page 65 note 1. Apiastrum meant also ‘balm a plant to which bees (apēs) are attracted.
page 65 note 2. By-form clūstellum Aldh., to clūslrum Not. Tir. 50.90 S; cf. mūscellum = mōs- tellim (1.26), where the variant spellings were influenced perhaps by analogues such as scrūtillus (: dialectal scrōtum = scrautum;see 7.19) and rūāus (: rodus— raudus). See Sommer2 79 Anm. end.
page 65 note 3. By-form crībrellum (cf. umbrella = umbellato umbra); cf. also M-L 2321.2 *garbellum, a contamination of crībellum and Arabic gerbal.
page 65 note 4. Probably a nonce-formation after the analogy of puer puellus; see Scaurus, GL 7.13.13 deminutionem… fratellus (= frāterculus, the proper formation on frāter, a consonant stem).
page 65 note 5. M-L 4864.
page 65 note 6. M-L 5010; masc, because of confusion with lĭbellus ‘ pamphlet ‘ (see List 1.20 and 21), not bilanx, which has fem, derivatives in Romance (M-L 1103).
page 65 note 7. By-form olīvastellum Gromat., to olīvaster; neut. after olīvum ‘ oil * or after other neut. diminutives in -stellum (canistellum plōstellum).
page 65 note 8. Neut. pīnastellum; see note 7.
page 66 note 9. Sacellus may have been made by inference from the noun saccllum or by the following analogy : sacrum : sacellum = sacer : sacellus.
page 66 note 10. See Not. Tir. 99.27 S for segestellum and cf. Varro LL 5.166 segestria quod fere stramenta erant e segete (but see Walde2 s. v. segestre); from the neut. pl. (apparently with sing, meaning) was made segestra -ae Edict. Imp. Diocl., and with reference to the latter and the gender of segestria the form segestrum CGL 2.181.36.
page 66 note 11. M-L 8659; neut. to terebrum (Hieron., -bra Cato) or after other diminutives ending in -bellum (crībellum dolābellum). The dissimilated form tenebellum is to be referred to tenebra ‘borer’ (M-L 8661.2).
page 66 note 12. Cf. the older dimin. tesserula Lucii.; by-form tessellum CGL 3.323.53 (neut. after pavimentum tessellātum Suet.) and perh. also tesellus (Gradenwitz, but apparently nowhere attested).
page 66 note 13. M-L 8965; a misformation like glomellum for *glomuscellum (cf. the correct formation tūberculum Cels.).
page 66 note 14. By-form umbrella CGL 3.326.62; cf. also CGL 3.370.35 libellum σκιάδιoν (contamination of scheda scida ‘ strip of papyrus, tablet ’ [: lĭbellas’] with σκιά ‘ shadow’, σκιάδιoν ‘ parasol’ ?). Friedrich, Demin. Bild. 60, would take umbella as a translation of Gr. σκιαδίσxη.
page 66 note 15. Plural like verenda -ōrum ‘ privates’ (Plin, sen.); the suffixal vocalism of verētilla was influenced perhaps by pēnicillus (for pēniculus in the meaning ‘ penis ’, see Plaut. Men. 286) or even by īlia -ium ‘ groin ’.
page 67 note 16. M-L 1365; neut. after cornū.
page 67 note 17. Petron. 74.13 (reading dubious); māchilla, not *māchella, because of macella ‘ stain ’ ? New formation : māchinula Hilar.
page 67 note 18. M-L 6143; with *pagina (M-L 6146)cf. Gr. πάγη ‘ snare’.
page 67 note 19. Originally an adjective (-īlla [sc. gemma] <-īno- + adjectival -lā-) with the meaning ‘precious stone of, i. e., found in Egypt’; cf. Friedrich, Demin. Bild. 66. Aegyptini is read by Lindsay for Plaut. Poen. 1291.
page 67 note 20. For the meaning cf. Colum. 2.10.20 cicer quod arietillum (codd., arietinum edd.) vocatur, and CGL 2.24.33 arietillum
‘leek’; mase, arietīllus occurs in Petronius (gender after ariēs?). The word was originally a neuter adjective (sc. cicer or ālium); see List 4.1, note 40.
page 67 note 21. Originally an adjective (sc. carō) or else made after the analogy of suīlla (sc. caro) ‘ pork’; neut. tergīllum (CGL 2.197.30 χοίρου δέρμα) was influenced by tergum or δέρμα. For *tergīnus cf. tergīnum ‘ leather whip ’.
page 67 note 22. Fem. after arcula ‘ perfume-box’; *unguīnus was shortened haplologically from *unguĭn-īmis, to unguen -inis.
page 68 note 1. M-L 118, to acucula M-L 120.
page 68 note 2. M-L 740; for -st- = -ss- cf. M-L 6442 *pestellum, to pessulum (influence of formations such as fistella, to fistula, and clōstellum, to clōstrum ?). Assula ‘ splinter, chip’ (*astula M-L 736.2) occurs in Plautus Merc. 130 at etiam asto ? at etiam cesso foribus facere hisce assulas ? (i.e., astutas, as a pun on asto ?). If assula may be referred, with Hofmann s.v., to asser ‘ pole the formation and the gender of the word nevertheless remain obscure (because of the disparity of the meanings ‘ pole ’ and ‘ splinter the diminutive of asser was reformed as a trisyllabic feminine after tabula ‘ board ’ ? Cf. asserculus ‘ small beam or pole ’ Colum., -um Cato).
page 68 note 3. Not. Tir. no. 15 S, to harcıda Not. Tir. no. 14 S; barca is first found in Paul. Nolan.
page 68 note 4. Neut. hotelliini M-L 1230 (neut. first in the meaning ‘ bowel ’ after intestina or viscera).
page 68 note 5. Gradenwitz, but apparently nowhere attested; the adj. is implicit however in Med. Latin brevicellum (DuCange).
page 68 note 6. CIL 4.5380.22.
page 68 note 7. By-forms : hucella CGL 3.340.25 (cf. οfella for offella), huccilla 3.7.39 (1 after frūstillum or ōscillum), bucilla 3.440.41, and buccellum 2.481.42 (neut. after frūstūm or ψωρίον).
page 68 note 8. Papy. Mar. 80.2.11 (564 A.D.); a var. MSS reading of these words is caca- (= cā-).
page 69 note 9. M-L 1612; fem. after plant-names (cf. nigella, maurella, etc.).
page 69 note 10. CIL 3.10107; the true reading of Pliny NH 36.178 is capitulum. Mase, capitellus CGL 3.589.21 (gender after abacus ?). See Friedrich, Demin. Bild. 59.
page 69 note 11. Neut. circellum CGL 3.441.77. Cf. circelliō, a derivative (or adaptation of circumcelliō) : CGL 5.596.41 circellio monachus per celias vagans; see Keller, Volksetym. 45.
page 69 note 12. Fem, clāvella ‘ peg, nail ’ M-L 1976 (contaminated with clāva clāvula ‘ cudgel; graft’).
page 69 note 13. Not. Tir. 92.3 S; for the spelling coe- = cē-, see Sommer2 78.
page 69 note 14. Conucella, like conuda (== coluc[u]la; M-L 2061), is a dissimilated form; see CGL 3.322.9, where either col- or con- maybe read. Conucella occurs as a plant- name in Diosc. (contaminated with cánula ‘ poley-germander Teucrium polium Linn.).
page 69 note 15. Medieval Latin; see DuCange s. v. (cotilla Gradenwitz, but apparently nowhere attested). Cōtella, like cōtula (-us) ‘ small stone ’ Anthim., was a misformation on the consonant stem cōt- (cf. cōticula Pliny sen.); cf. perhaps facella, to fax -cis.
page 69 note 16. Gl. Paris, ap. Quicherat, Addenda Lexicis Latinis 62 (Paris, 1862).
page 69 note 17. Mase. Pliny sen. (indeterminate Colum. 12.7.1-i cauliculum); cf. digitillum Pliny sen. (i by vowel-assimilation?). See Friedrich, Demin. Bild. 47 and 59; for the semantics cf. perhaps *cantharella ‘ mushroom ’, to cantharulus ‘ tankard’.
page 70 note 18. Medieval Latin; see DuCange s.v. and cf. M-L 2737 dominicellus. For dominus in the meaning ‘ master’s son, young man see Plaut. Capt. prolog. 18 and Cato Agr. 1.4.
page 70 note 19. M-L 2969. Excipulum (with primary -elo-) is found in Pliny sen.
page 70 note 20. Not. Tir. 112.84 S, to fascinābulum Not. Tir. 112.83 S.
page 70 note 21. M-L 3276.
page 70 note 22. Neut. flōscellum Itala Exod. 38.16 (translating
‘ wreath ’ of the Septuagint).
page 70 note 23. M-L 3418.
page 70 note 24. Mase, formellus ‘ kind of cheese ’ Gildas Sapiens (gender after cāseus).
page 70 note 25. M-L 3615.
page 70 note 26. M-L 3791.
page 70 note 27. Cf. mase. *glomellus M-L 3797, to glomulus M-L 3800 (both masc. after globus). Globus and glomus -eris were confused : cf. *globuscellum (M-L 3794) for globellus and glomellum for *glomuscellum (cf. perhaps 14.3 *astella).
page 70 note 28. Medieval Latin; see DuCange s.v.
page 70 note 29. M-L 4640, to iuvenculus ‘ young bullock ’ Vulg.
page 70 note 30. M-L 4881, to lancula (languid) ‘ balance(-pan) ’ Varro (cf. lancicula ‘ small dish ‘ Arnob.); for the formation cf. facella (facula fax) and nucella (nucula nux).
page 70 note 31. Not. Tir. 40.82 S.
page 71 note 32. M-L 5305.
page 71 note 33. Var. MSS reading marcellus; martellus in Romance (M-L 5379; t for c because of analogy to Mārtius ‘ martial ’ or by hyperurbanism; cf. the reverse in mūscellum == mōstellum). Neut. martellim CGL 3.23.22, to marclum 3.204.18 (neut. after other instrumentalia in -culum). Back-formation : marcus Isid. and Romance (M-L 5348); cf. CGL 5.572.41 marcus fortior malleus, marculus diminutivum, martellus tnediocris malleus. Marculus is from *mal-tlo-s > with dissimilation *marklos > with anaptyxis marculus (see Walde2 s.v. marcus; but marcellus is not to be derived phonetically from *marklo-lo-s, since phonetic development is precluded by the late appearance of the word).
page 71 note 34. See Forcellini s.v. for attestation.
page 71 note 35. Cf. CGL 3.205.28 muscellus
(which apparently never means ‘ boat’ in Greek literature) and Isid. 19.1 musculuscurium navigium (apparently a translation of Gr.
‘ boat’, which was taken for μυίδιον ‘ mouse’). See Friedrich, Demin. Bild. 59 and 72.
page 71 note 36. Var. MS reading naucella, to naucula Pliny iun.; cf. the doublets avicella aucella (List 6.5).
page 71 note 37. M-L 5943.
page 71 note 38. By-form : CGL 3.316.17 nucillae
άρυα ποντι
ά (‘filberts’); for the i, cf. 14.87 bucella/bucilla, and for the formation cf. facella to facula fax (nucicla occurs in Isid.).
page 71 note 39. By-form offella (also in Romance : M-L 6042); cf. the doublets buccella bucella and see Sommer2 206.
page 71 note 40. M-L 6200.
page 71 note 41. M-L 6257.
page 71 note 42. M-L 6349 (reading of Petron. 57 is dubious).
page 71 note 43. M-L 6442; for -st- = -ss-, cf. *astella to assula and the doublets pustula pussula.
page 72 note 44. Not. Tir. 105.85 S.
page 72 note 45. CGL 2.154.18.
page 72 note 46. CGL 3.632.69 prunella sive prunella; cf. 3.536.34 progella and 3.591.44 bumella. The feminine in these corrupt glosses is the converted neut. pl. of *prūnel- lum (cf. M-L 6798 *prūna = prūnum).
page 72 note 47. Gradenwitz, but not in the lexica; the ‘ double ’ formation is implicit however in the ‘ triple ’ puericellulus (see DuCange s.v. puericelluli).
page 72 note 48. M-L 7036 and 7037; neut. after bracchium ‘ bough ’ or rāmālia ‘ sticks ’. For the formation cf. *fāgustellum, to fāgus -ī. The ending -tellum is for -cellum by hyperurbanism (see note 33 supra) or was influenced perhaps by olīvastellum (to olīvaster). Rāmusculus occurs in Hieronymus.
page 72 note 49. M-L 7471.
page 72 note 50. M-L 7481.
page 72 note 51. By-form sacellus (Sommer2 206) CGL 3.626.50; neut. saccellum CGL 3.194. 34 (gender after βαλλάντιον in gloss or marsūpium ‘ purse ’).
page 72 note 52. Medieval Latin; see DuCange s.v.
page 72 note 53. M-L 7597. Sarculum is from *sar-tlom, to sartō -īre ‘ weed’; see Walde2 s. v.
page 72 note 54. M-L 7749. See Walde2 s.vv. scrōfa and scrōfula.
page 72 note 55. Not to spīcula ‘ ground-pine ’; spīculum became neuter perhaps after iaculum.
page 72 note 56. M-L 8356.
page 73 note 57. Mase. M-L 8472; cf. neut. surcellum Apic. (gender after bracchium ‘ bough ’), to surculum Venant., and the further by-form surcillus Plin. Valer. Surcellus has no precise morphological parallel (cf. perh. puer cuius; or were sur cellus sur cuius induced by *sarcellum sarculum ?).
page 73 note 58. M-L 8614, to tegula M-L 8613.2 (= teganum < Gr. τάγηνον ‘ fryingpan’).
page 73 note 59. M-L 8771 (neut. pl. as fem. ?).
page 73 note 60. Var. MSS reading unguella. For the formation ungula Ennius (cf. unguiculas ‘ finger-nail’ Plaut.), cf. perh. cōtula = cōlicula; is it possible that ungida was felt to be the diminutive of uncus ‘ hook, hooked ’ ? Ungula acquired fem, gender in order to avoid homonymity with ungulus ‘ finger-ring ’.
page 73 note 61. M-L 9123, to vāgīnula ‘ husk ’ Pliny sen.
page 73 note 62. M-L 9140; the meaning ‘ wohl nach der Bewegung des Federbusches, die mit dem Schütteln eines Siebes (vannulus 9143) verglichen wird. ’
page 73 note 63. CGL 2.433.14; cf. vascella 3.22.41 (fem, after σκευ
gloss or perh. neut. pl. as fem.).
page 73 note 64. CGL 3.14.49; also Romance (M-L 9208).
page 73 note 65. M-L 9363.
page 73 note 66. Not. Tir. 96.32 S.
page 73 note 67. See Forcellini s.v. (Onom. Lat. Gr. focillus
; not in CGL).
page 73 note 68. See Forcellini s.vv. for attestation.
page 74 note 69. See Walde2 s.v. (a derivative of portilla). Porcillāca (also porcilāca) was perhaps influenced by porc(ellus) or porcilia ‘ sow’.
page 74 note 70. The ū (long before -nct-) became short in Late Latin by influence of pŭngō; see M-L 6847.
page 74 note 71. To verticulus ‘ whorl’ M-L 9255; cf. also *vertibellum ‘ basket for yarn’ M-L 9251, which was made with reference to verticulus or verticillus. Cf. also CGL 3.322.5 verticillus σφόνδυλoς ‘vertebra’ (3.366.45 verticellus), to verticulus ‘joint’ Solin., and CGL 2.435.58 verticillum σπόνδυλος (3.21.8 verticellum), to verticulum ‘ joint ’ Cael. Aurel.; these words are all to be referred to verticula Lucil. = vertebra, the gender of which was influenced by that of corresponding Greek terms (cf. σφόνδυλος,
,
‘ hip-joint’).
page 74 note 72. Cf. the misformation cōdicilla ‘ tail ’ Apic. (as though to *cōdex -icis ‘ *tail ’).
page 74 note 73. See Hofmann s.v. cillō and cf. M-L 1914 *cillicare ‘bewegen’.
page 74 note 74. With the exception of a handful of first occurrences in Celsus, Columella, Petronius, Juvenal, Martial, Apuleius, and some inscriptions of uncertain date, the majority of the diminutives in -ello/ā- of List 14 were formed apparently during and after the third Christian century. See Baehrens, Sprach. Komment, z. vulgärl. App.
Probi 121-2, who reaches much the same conclusion and speculates on the reasons for the comparative restriction of analogous formations during the earlier period of the language.
page 75 note 75. The formations cōtella (for *cōtteella), facella, glomellum, lancella, nucella, surcellus, and ungella have analogues in the Republican avillus (if to ovis), fracillī, mellilla, and murmurillum. Nucella perhaps ousted the less easy *nucicella; for glomellum, surcellus, and ungella, see notes 27, 57, and 60 supra.
page 76 note 1. M-L 322.
page 76 note 2. Thes. Nov. Latin, (ed. Mai) 48.
page 76 note 3. M-L 709; fem. after other botanical terms (lapella, malvella).
page 76 note 4. For the meaning ‘ * shawm ’ see M-L 1484.2.
page 76 note 5. Mase, in Romance (M-L 1561); neut. -um after arvum or territōrium.
page 76 note 6. M-L 1644, to cappa Isid. and Romance (M-L 1642).
page 76 note 7. CGL s. 522.21 cappellus galerus; mase, after galērus or pilleus.
page 76 note 8. M-L 1731, to CGL 2.571.34 cascabus caccabus grandis; cascabus was perhaps a contamination of cassus ‘empty, hollow ’ (or cassis ‘ helmet ’) and caccabus ‘poť. For the semantics of ‘ pot ‘ > ‘ bell cf. the development of campāna (see Hofmann s v.).
page 76 note 9. M-L 1772; see Hofmann s. v. caucum.
page 76 note 10. Medieval Latin; cf. DuCange s.v. cavellus ‘a kind of tribute money’, and see Hofmann s. v. caulae.
page 76 note 11. M-L 1916; fem. as a ‘ Kosewort für kleine Kinder’?
page 77 note 12. M-L 1934, with the meaning ‘ horse-collar, hame ’ in Romance; with *cipellus, cf. of ella for offella.
page 77 note 13. M-L 1474.2.
page 77 note 14. Medieval Latin; see DuCange s. v. and cf. M-L 4747 kotta (Frankish) ‘ garment ’
page 77 note 15. Medieval Latin; see DuCange s.vv. cotarelli and cotherius.
page 77 note 16. M-L 2315 (> Fr. crécelle ‘ rattle’); perh. also to crepō, with the ending -icella derived from diminutives of i-stems (particella; cf. also vert-icillus, as though to vertō).
page 77 note 17. Medieval Latin; see DuCange s.v. and cf. dominellus (14.27) and its note.
page 77 note 18. M-L 2737; not to dominica (-us) ‘ Sunday ’ M-L 2738. For the formation cf. puericellus, to puer.
page 77 note 19. Exagella is fem, after lībra or lancula.
page 77 note 20. M-L 3141; the formation is curious : perhaps the ending -anello/ā- detached itself by misdivision from raphanella ‘ radish ’ (to raphanus), or was the result of the contamination of -ăbello/ā- (cascăbellus) with the ending -nello/ā- (campānella, dominellus; cf. also *portīnellum, to porta).
page 77 note 21. M-L 3146; for the formation cf. *rāmustellum (14.66), to rāmus -ī. The ending -stellum was induced perhaps by olīvastellum ‘ olive-tree ’, to olīvaster.
page 77 note 22. M-L 3299; the word is fem, perhaps after fimbria ‘ fringe ’ or tēlae ‘ warp- threads ’.
page 77 note 23. M-L 3381; a misformation (cf. 14.33 floscellus) assisted perhaps by calic-ellus, to calix calic-is, or, equally likely, -iscellus was a contamination of Gr. -
- and Lat. -ello-.
page 77 note 24. A misformation; cf. perhaps baticilla for *batidicilla, to batis -idis. See Thes. Ling. Lat. s. v., where emendation to fluvicellum is proposed.
page 78 note 25. M-L 3728 (> OFr. genevelle ‘ door-hinges ’); for the formation cf. *manabel-al to manus : the ending -abella (ă
) was perhaps a contamination of the dimin. -instru, mental suffix seen in tabella and sūbella with the ending -ăbello- of caccăbellus (to caccăbus).
page 78 note 26. M-L 3794; for the formation see 14.38 glomellum and its note.
page 78 note 27. M-L 3799; for the formation cf. perhaps *flōriscellus (see note 23 supra).
page 78 note 28. Cynoglossum officinale; see Isid. 17.9.66 (lappella cod. D) and cf. CGL 3.537.29 cinglosa id est lingua canina id est lappella. See Sofer, Glotta 16.5. With the relation of lapella to lappāgō ([lappa) cf. perh. arilla(tor), to arrabō (arra; see 9.1 and note).
page 78 note 29. See Sofer, Glotta 16.42.
page 78 note 30. M-L 5280; for the formation cf. *genabella (note 25 supra), and manibul ‘ plow-handle ’ Varro.
page 78 note 31. M-L 5302.
page 78 note 32. M-L 5352; for the formation cf. flūmicellum for *flūminicellum.
page 78 note 33. CGL 3.630.40, referred incorrectly by Meyer-Lübke (5680b) to maurus ‘Moor, black probably not to be equated with CGL 3.629.35 scamaria ia id est morella id est aquatica (3.592.52 maurella herba ficaria) : see CGL vol. 6 s. v. maura and vol. 7 s.v. uva lupina.
page 78 note 34. Medieval Latin; see DuCange s.v. morsellus (morsillus).
page 78 note 35. M-L 5767, to *mūsus ‘mule ‘ (M-L 5784); the formation was influenced by mŭscella ‘fly’ and mūscellus ‘small boat’. Morphologically, cf. surcellus (14.78), to sur us -ī.
page 78 note 36. M-L 5802, to myrta (mur-) M-L 5801.
page 78 note 37. M-L 6144; fem, after other nouns denoting measures.
page 78 note 38. M-L 6193 a.
page 79 note 39. Masc. after nucleus or conus.
page 79 note 40. M-L 6534.
page 79 note 41. Thes. Nov. Latin, (ed. Mai) 473.
page 79 note 42. M-L 6676; for the ending cf. note 20 supra : *portīnellum apparently owes its formation partially to a contamination of postīcum ‘ postern ’ with portella (14.62).
page 79 note 43. M-L 6819; cf. pullus and pullulus ‘ dove, chick ’ as terms of endearment.
page 79 note 44. Medieval Latin; see DuCange s.v. poledrus ‘pullus equinus ’ and cf. M-L 6825 *pulliter, *pulletru, pullitru ‘ Füllen ’ (contaminated with pullus ‘foal’).
page 79 note 45. M-L 6851.
page 79 note 46. M-L 7050; fem. after rādīx.
page 79 note 47. M-L 7255 a; fem. after plaga ‘ hunting-net ’ or after rēte, which was sometimes feminine (Varro). For the formation cf. brevellus.
page 79 note 48. M-L 7823.
page 79 note 49. A nonce-formation; trisyllabic after suīllus ‘ of swine’ or comparable to brevellus.
page 79 note 50. Var. MSS reading verbella(d. M-L 9270); fem. after ovis. Not to be derived with Wölffin, ALL 12.303, from *vervēc’la (-kl- remained or developed an anaptyctic vowel : Sommer2 228), but to be reckoned as an analogical formation like suellus. See CGL 2.416.24= GL 1.553.28.
page 79 note 51. Trisyllabic after iūscellum ‘broth’ (14.41) or supplanting a difficult *viscuscellum.
page 79 note 52. During the Imperial period -cello/ā- apparently attached to nominatives in -s (see§ 45 supra) and -cillo- to nominatives in -r (agercillus and 15.2 corcillum).
page 80 note 53. See Löwe, Prodr. 432, and CGL 6.17 s.v. acilli (arilli cod. and Romance : M-L 646; acini coni. Bücheier).
page 80 note 54. M-L 990; for *batidicilla (unlike the phonetic lapillus < *lapid-los).
page 80 note 55. The loss of l in būgillō would then be dissimilatory; for the formation cf. lapella (16.33), to tappāgō. The meaning ‘ground-pine ‘ (= abiga, an herb used to effect abortion) may be arbitrary, inasmuch as būgillō might have been associated with bulga ‘bag’, in the sermo castrensis also ‘ belly, womb’; cf. Hofmann s.vv. būgillō and bulga.
page 80 note 56. Cf. Paucker, KZ 23.178 (List C. 27) and M-L 3504 *frīcta ‘ fritter’; the loss of r in f(r)ītilla was dissimilatory; for -t- = -ct-, cf. fīttila (18.36) = fīctīle. Otherwise Walde2 s.v. (to fingō, in view of Goth, daigs ‘ dough ’).
page 80 note 57. CGL 5.599.43 fractillum ad frangendum piper(= lapis Albānus ‘ peperino’ ?); cf. frāctāria -ac ‘ machine for crushing flint ’ Pliny sen.
page 80 note 58. CGL 2.334.3 frittila (frutilla cod. e, after frutex ‘ bush ’ ?)
; for the formation cf. fringilla (9.6).
page 80 note 59. Neut. fritillum CGL 2.73.44 and 5.457.52 (gender after pyrgum ‘ dice-box ‘ in gloss); Walde2 s.v. refers fritillus to fritinniō ‘ chirp’.
page 80 note 60. For the formation cf. *margella (16.37), to margō -inis.
page 80 note 61. Attested by Hofmann s.v. cavilla.
page 80 note 62. Cf. CGL 4.153.8 purpurilla locus extra portam, ubi scorta prostrant : dictum est autem vocabulum quod matronae stola, libertinae toga, prostitutae purpurea veste uterentur; see Būcheier, ALL 2.117, and cf. List 18.67 turtuvilla.
page 80 note 63. A nonce-word: Apul. Meta. 8.2 (ed. Vliet; Teubner, 1897). It id conceivable also that tempestillus (ī) was made on a *tempestīnus (cf. 4.12 rēgtīlus).
page 80 note 64. Cf. M-L 9010.2 *lurturella; for the formation cf. 9.11 murmurillum.
page 81 note 65. Or perhaps vespillō was a euphemism suggestive of vesper * evening’(d. 10.28 tranquillas) : Isid. 20.10.5 unde et permansit ut mortuos faces antecedant; but see Walde2 s.v. and Muller, Altit. Wörterb. 541, on this puzzling word. Keller, Lat. Etym. 127-8, thinks the word an adaptation of vespertīliō ‘ bat’, as ‘ is qui vespertino potissimum tempore conspicitur vel munere suo fungitur. ‘
page 81 note 66. See Forcellini s.v. for attestation; -illa was induced by analogy to proper names in -illo/ā- (cf. 15.4 nepōtillus -a).
page 81 note 67. I have purposely excluded from this count the ‘ single ’ -ello/ ā- and -
llo/ā- of r- and n-stems as not being illustrative of the grafting of these endings as independent elements on stems of a great variety (see Lists 8, 9, 16 and 17). An indiscriminate count of all instances of the ‘ single ’ suffix in both periods of the language totals 223 examples of the latter as against 139 instances of the ‘ double ’ suffix.
page 82 note 1. See Isid. 20.16.4, CGL 5.589.22, and Hofmann s.v. antilēna.
page 82 note 2. M-L 636 (cf. 10.3 argilla ‘ potter’s clay ’ ?).
page 82 note 3. M-L 766.2.
page 82 note 4. M-L 864.
page 82 note 5. See DuCange s.v. bargilla and cf. CGL 4.115.8 mantica bargila.
page 82 note 6. CGL 5. 347·12; a contamination of bǒvīe, bū
īle, and būbellus ‘ of oxen ’; for hovilla see 4.5 bovīllus and its note.
page 82 note 7. Var. form brabyla; see 10.16 murmillō and its note.
page 82 note 8. M-L 1602.
page 82 note 9. M-L 1686.
page 82 note 10. See Du Cange s.v. 1. carnellus (carrus plus -nello- as in 16.48 *portīnellum??).
page 83 note 11. See DuCange s.v. 2. carnellus.
page 83 note 12. CGL 5. 354.9 (corbellus coni. Götz).
page 83 note 13. See CGL 5.349.41 and cf. M-L 1722 cartallus.
page 83 note 14. Cf. Isid. 19.1.25 Lind, caupulus (varr. -ilus and -illus) and CGL 5.614.13 caupilus. See Hofmann s.v. caupulus and cf. M-L 1780 caupulus.
page 83 note 15. See Hofmann s.v. cīlō.
page 83 note 16. See Keller, Volksetym. 54.
page 83 note 17. See CGL 5.617.43; constructed from cro(cus) ‘ saffron ’ plus the ending -stello/ā- seen in some botanical terms (oleastellus, pīnastellus, etc.) ??
page 83 note 18. See Leumann 172. IV.D.
page 83 note 19. CGL 2.479.31 fitilla olla (does -illa represent pl. -īlia or -īle with expressive gemination and gender after olla ?).
page 83 note 20. M-L 3626.
page 83 note 21. See Anthologia Latina (ed. Riese) 117.16 and Niedermann, Ĕ und Ĭ 65; for the final syllable cf. perh. matelliō ‘ pot’, to 6.14 matella.
page 84 note 22. Not to gīrus (g
rus) ‘ circle’ with Isidore: Orig. 20.15. 2 Lind. : girgillus quod in giro vertatur: est enim lignum in transversa pertica mobile ex quo funis cum situla vel utre in puteum dimittitur auriendae aquae causa; see Cuny, MSL 19.199, and Sofer, Glotta 16.40-1. The original *galgillus > with dissimilation *gargillus > girgillus through popular association with gīrus.
page 84 note 23. See Arnob. 7.230; gender after placenta -ae, which was popularly referred to placēns( : piacēre;see Keller, Volksetym. 83), as Gr. άρεστήρ was to άρέσxω ‘conciliate’.
page 84 note 24. Deuerling’s emendation seems very probably correct : gravastellas graves, id est annis; see 1.16 grāvastellus.
page 84 note 25. See Leumann 176. II. A. 2, and cf. rubelliō ‘ a kind of fish ’.
page 84 note 26. See CGL 2.353.36 and cf. M-L 5353.
page 84 note 27. See Walde2 s.v. metella.
page 84 note 28. Cf. CGL 2.373.29 mascella μουλάριον (cf. 5.298.22); the result of the contamination of Gr. µύλη ‘ mill; knee-pan ’ with mūlus = *mūsus (M-L 5784) ‘ mule’ ??
page 84 note 29. See Walde2 s.v. nītēdula.
page 84 note 30. M-L 6230.
page 85 note 31. M-L 7801.
page 85 note 32. Cf. CGL 5.612.42 turturilla ita dictus locus in quo corruptelae fiebant, quod ibi turturi opera daretur, id est panem (z. e. peni), and see Nettleship, Journal of Philology 20.187 (tuhmilla, to Tūtūnus ‘ Priapus ’, proposed, in view of a var reading tuturilla). Cf. 17.12 purpurilla.
page 85 note 33. Read perhaps velum peplum with Landgraf, ALL 9.442.
page 86 note 1. This fact squares with one's impression that at least in French the suffix -ello/Ā- has many more representatives than the suffix -illo/Ā- (types chapelle < cappella, chapeau < cappellus, and cheveu < capillus). For the endings -ello/Ā- and -cello/Ā- in Romanic languages, see W. Meyer-Lübke, Grammaire des Langues Romanes 2. 591-4 and 615-7 (Paris, 1895).
page 87 note 1. See the article of A. Funck, ALL 4.68 sqq. and 223 sqq. For the type focillō — focilō foculō, see J. Samuelssohn, Verba auf -ŭlāre, -ĭlāre, Glotta 6.225 sqq.
page 87 note 2. M-L628.
page 87 note 3. With haplological loss of -tī- (cf. perhaps dentiō beside dentī tiō i teething ’); see Hofmann s.v.
page 87 note 4. Cf. cucurrīre ‘ crow’ and see Hofmann s.v. cacabō ‘ cackle ’ (of the partridge); see also note 13 infra.
page 87 note 5. See note 1.
page 87 note 6. ‘ Misformed’ since capulō was not a diminutive formation; cf. concipilō ‘ cut in small pieces’ (Plaut. Truc. 621). See Hofmann s.v. cāpō ‘ capon’ and note 1.
page 87 note 7. See Funck, ALL 4.244, Hofmann s. v. cillō, and M-L 1914 *cillicāre ‘ bewegen’; cf. Festus 194.8 M oscillum (i. e. a ritual-mask hung up in the open air, on trees, etc.) Santra dici ait, quod os cillant (oscellant Lind.), id est inclinent praecipitesque afferantur.
page 87 note 8. Cf. the PPP capillātus ‘ having hair ’.
page 88 note 9. See Plaut. Mil. 118o, Muller, Altit. Wörterb. 503, and Hofmann s.v. Cf. faber ‘ joiner, carpenter ’.
page 88 note 10. See CGL2.69.52; for ā = au , cf. CIL 6.667 Cladivs (ā ) = Claudius.
page 88 note 11. See Walde2 s.v.; -ll-in Romance (M-L 3237).
page 88 note 12. See Paul.-Fest. 508.13 Lind. (— 371.3 M).
page 88 note 13. See Anthologia Latina (ed. Riese) 762.25.
page 88 note 14. M-L 5199.
page 88 note 15. Hardly to strigilis ‘scraper’ (ALL 4.229); for the omission of the weak sound n in writing, see Sommer2 238. Derivative : obslrigillus -ī ‘ shoe-sole, sandal ’.
page 88 note 16. See Walde2 s.v. and Sommer2 502.
page 88 note 17. M-L 7641.
page 88 note 18. Sigillum had the meaning ‘ seal ’ especially in late Popular Latin (see M-L 7904; cf. also Hor. Ep. 1.20.3).
page 89 note 19. See Pompeius, GL 5.220.35, and Ernout-Meillet s.v. sūgō (whence sūgillō, with a transitional meaning ‘ *make a mark by sucking ’); Muller, Altit. Wörterb. 437-8 (akin to Goth, sinks ‘ ill’, NE sick); and Sommer2 2632 after Prellwitz, KZ 45.71 (< *subs-gillō ‘ blaue Flecken schlagen ’, to Lettish zils ‘ blau ’ ??).
page 89 note 20. Derivative : tītillus ‘ tickling; armpit ’ (Late Latin; cf. M-L 8757).
page 89 note 21. See Muller, Altit. Wörterb. 516, and Hofmann s.v. convexus (*
ak-ro-la
ō > *vacellō > vacillō by some analogy ?); to root *
a-n-q- ‘ be bent’. For vacc- see Lucr. 3.504 (vac- 1.806) and Sommer2 207 Anm. 2 end (vac[c]illō, with a meaning ‘ *waddle like a cow popularly associated with vacca ??). See also Pompeius, GL 5.221.10 sqq. (connection with vagor disallowed), and Funck, ALL 4.236-7.
page 89 note 22. See Walde2 s.v.
page 89 note 23. See § 22 supra, note 4.
page 90 note 24. See Sommer2 § 118.2.