Understanding 'percepset' instead of 'percepisset'

Question

When looking at the L&S entry for percipio, I came across a surprising perfect form: percepset. The contraction percepisset > percepset lookst similar to cogitavisset > cogitasset. Are contractions similar to percepset common for verbs whose perfect stem does not end in V? I do not recall ever seeing any. If such forms are found, is it specific to some eras, verbs, forms, or contexts?

Answer 1

I believe they are very rare, but it is possible to find some others. I have found vixet in the Aeneid:

[Publius Vergilius Maro, Aeneis 11.118]

apparat, his mecum decuit concurrere telis:
uixet cui uitam deus aut sua dextra dedisset.
nunc ite et miseris supponite ciuibus ignem.'

On which comments Maurus Servius Honoratus, a late fourth-century and early fifth-century grammarian:

[Maurus Servius Honoratus, In Vergilii Aeneidos Libros:]

11.118.1 vixet pro 'vixisset': et est syncope.

3.152.1 INSERTAS aut clatratas; aut non seratas, ut sit quasi 'inseratas', id est non clausas, et dictum, quomodo 'asprosque molares' pro 'asperos', conpostus' pro 'conpositus', 'vixet' pro 'vixisset'. {⁴³vel 'insertas fenestras' quas lumine suo luna inseruerat, ab inserendo, quod se per rimas insereret.

So he considers it simply a syncope. As I can only find any forms on -ixet in the Aeneid, it may have been invented by Virgil metri causa, but possibly with some licence based not only on poetry but also on archaic variants, when morphology had been less fixed.

I can find -pset only in two sources (one of whom is Cicero):

[Marcus Tullius Cicero, De Officiis 3.98.9]

Quod ni Palamedi perspicax prudentia
Istius percepset malitiosam audaciam
Fide sacratae ius perpetuo falleret."

[Anonymi Comici et Tragici, Tragoediae Poetarum Incertorum 59]

Quod ní Palamedi pérspicax prudéntia
Istíus percepset málitiosam audáciam,
Fidé sacratae iús perpetuo fálleret.

Answer 2

András Cser's "The nature of phonological conditioning in Latin inflectional allomorphy" seems to say contraction like this occurs somewhat often in the perfect forms of verbs that have perfect stems ending in /s/ (as in Cerberus's example vixet ~ vixisset), and gives a few examples. Cser doesn't mention any patterns of variation, but he does offer a phonological analysis:

2.3.2.2. Vowel deletion after [s]

One minor point of complication involves s-final perfective stems (excluding -ls-, -rs-) combining with is-affixes. In more than just a handful of cases the vowelless affix variant appears, and the number of adjacent [s]’s is reduced. Thus derexisti ∼ derexti ‘arrange’ PresPerf2Sing, divisisse ∼ divisse ‘divide’ PerfInf, accessistis ∼ accestis ‘approach’ PresPerf2Plur, admisisse ∼ admisse ‘send to’ PerfInf, and many others. On the face of it this looks like the loss of an [is] sequence specifically after [s] in these particular constructions (as is the tradition, see Leumann 1977, 598), but it is more economical to analyse the disappearance of the vowel as being the same allomorphic variation as that seen after vowel-final stems, and the disappearance of [s] as resulting from an independently motivated phonological process of degemination, because in this case we do not need any extra processes – apart from stipulating the somewhat odd context [s] for the otherwise postvocalic morpheme variants.

(p. 13)

In Latin Hexameter Verse: An Aid to Composition, by Samuel Edward Winbolt (1903), I found mention of the forms vixti (for vixisti), extinxem (for extinxissem), and finxe (for finxisse) (p. 212).

I can't find any examples of derexet for derexisset, though.

Other sources that seem to have relevant information:

Ancient forms of a future perfect in so, a perfect and pluperfect subjunctive in sim and sem, and a perfect infinitive in se sometimes occur. They may, in general, be formed by adding these terminations to the second root of the verb; as, recepso, emissim, ausim from the obsolete perfect, ausi, from audeo, confexim and promissem: divisse and promisse. But when the root ends in x, and frequently when it ends in s, only o, im, em, and e, etc. are added; as, jusso, dixis; intellexes, percepset; surrexe, sumse. V, at the end of the root, in the first conjugation, is changed into s; as, levasso, locassim. U, at the end of the root, in the second conjugation, is changed into es; as, habesso, licessit. Sometimes the vowel of the present is retained in these forms, though changed in the other parts derived from the second root; as, capso, faxo (facso), faxim (facsim).

(A grammar of the Latin language; for the use of schools and colleges, by E.A. Andrews and S. Stoddard, 1857, p. 122)

Some similar-looking forms seem to have been used as archaic future subjunctives:

[A "future pluperfect" tense] occurs in Virgil, Aenid, 11, 467, Jusso, and was found in Cicero, de Legg. 2, 9, Jussit, until altered to Jusserit by modern editors.--Not satisfied with Vossius'es formation of it from the future in ERO, I derive it from the (contracted*) pluperfect subjunctive, as Ama'ssem, Amasso -- Summo'ssem, Sumosso -- Recep'sem, Recepso -- Effec'sem, or Effexem, Effexo -- Jus'sem, Jusso -- Audi'ssem, Audisso. -- The verbs in UI took ESSO, as Habesso, in Cicero, de Legg. 2, 8.

(Latin prosody made easy, by John Carey, 1819, pp. 93-94)

On p. 220, Carey gives a list of a number of other forms "all found in classic authors" that show this kind of "syncope of IS, ISS, or SIS"; I have rearranged them into groups by their endings, and bolded the ones where the standard perfect stem doesn't seem to end in /s/:

• -sti (active perf. ind. 2s):
  Scripsti, Conscripsti, Praescripsti, Subrepsti, Consumpsti, Cœpsti, Mansti, Sensti, Misti, Promisti, Amisti, Dixti, Intellexti, Advexti, Prospexti, Aspexti, Luxti, Abduxti, Adduxti, Induxti, Subduxti, Instruxti, Depinxti, Devinxti, Emunxti, Immersti, Tersti, Exclusti, Revixti, Exstinxti, Percusti

• -stis (active perf. ind. 2p)
  Accestis

• -se (active perf. inf.):
  Scripse, Carpse, Sumpse, Consumpse, Cepse, Promisse, Elisse, Divisse, Admisse, Decesse, Dixe, Illuxe, Illexe, Advexe, Circumspexe, Surrexe, Abstraxe, Prospexe, Despexe

• -sem (active pluperf. subj. 1s):
  Conclussem, Faxem, Interdixem, Exstinxem

• -ses (active pluperf. subj. 2s):
  Intellexes

• -set (active pluperf. subj. 3s):
  Percepset, Recesset

• -semus (active pluperf. subj. 1p):
  Erepsemus

Note that despite what Cser says about -rs- being excluded, Carey mentions some contracted words ending in -rsti.

The Art of Latin Poetry (1828) specifically mentions "omitting the syllable si before sti" as a "contraction chiefly used by the comic writers ; but Vergil has accestis, for accessistis, Æn. i 205; and Horace evasti for evasisti, Sat. ii. 7. 68" and gives a separate description of "rejecting is or iss , after x,; Direxti, for direxisti, Virg. Æn. vi. 57. Extinxtem, for extinxissem, Id. Æn. iv. 606. Surrexe, for surrexisse, Hor. Sat. i. 9. 73" (p. 41).