In the Proto-Italic language, verb conjugations were a complex system of inflectional morphology with conjugation according to mood, voice, tense, person, and number.
Present The South Picene form possibly reflects a second-person plural present active indicative ending , which is incongruent with the Latin ending . Alternatively, the South Picene form has been interpreted as an imperative or a second-person preterite form. Nevertheless, accepting the view that the term is indeed a second-person plural form, Weiss suggests that Proto-Italic perhaps possessed the primary second-person dual and plural endings and alongside the secondary second-person dual and plural endings and . Weiss argues that Sabellic generalized the dual forms and leveled the from the secondary to the primary endings, whereas Latin continued the original plural forms. There is evidence of a similar development in Umbrian, where the second-person plural imperative ending may reflect an original Proto-Indo-European dual form , which could imply that Umbrian also generalized the dual for the plural. Paelignian, a Sabellic language of the Oscan group, contains the second-person plural imperative form (“go”), which could display the inherited PIE second-person plural ending , although it is also possible that this specific form was borrowed from Latin.
The Proto-Indo-European mediopassive second-person singular ending perhaps shifted to under the influence of the second-person singular active. Afterwards, the pre-form would have rhotacized into Latin passive ending . Alternatively, the linguist Kazuhiko Yoshida reconstructs for Proto-Indo-European a second-person singular mediopassive ending , which he suggests—in Italic—evolved into , which then shifted to under the influence of the second-person singular ending. The alternative Latin passive ending perhaps emerged as the result of later enlargement with the PIE athematic active ending . In some Latin dialects, the remodeling occurred prior to the weakening of word-final , producing a proto-form , yielding , as reflected in the form ("walk about"). Both Sabellic and Latin also innovated second-person plural passive forms marked by a base , as reflected in the Latin ending . The historical development of this form is unclear, though it may derive from the nominative plural of a Proto-Indo-European mediopassive participle , which perhaps yielded Proto-Italic , whence perhaps Latin .
In the third-person plural, Latin has generalized the original thematic ending , even to originally athematic forms such as (< PIE ), whereas Sabellic generalized the athematic ending , such as in Oscan . Sabellic perhaps innovated the new third-person mediopassive ending , which is continued in Umbrian as the endings and . South Picene (“is erected?”), however, may show a similar development of the third-person singular passive as Latin. Nevertheless, in other Sabellic languages, the new ending was utilized in the primary tenses, whereas the original ending was restricted to secondary forms. Thus, endings such and , persist in Umbrian, but only in the passive subjunctive. According to the linguist James Clackson, Proto-Sabellic perhaps developed separate primary and secondary third-person passive endings and , the latter of which was secondarily generalized to all passive forms in South Picene. Clackson proposes that, in contrast, the Oscan group of languages generalized the Proto-Sabellic primary forms to the entire passive paradigm. Hence, Paelignian reveals terms such as the imperfect subjunctive (“might be built”). Regarding the ultimate origin of these new endings, One theory suggests that the they perhaps emerged on the model of the third-person form . Alternatively, it has been proposed that the Sabellic forms reflect innovative , itself possibly adapted to the vocalism of the first-person singular passive ending .
Imperfect To mark the imperfect, a new past indicative suffix of was created in the Italic languages. This morpheme is continued in Latin terms such as and possibly by the Oscan form ("they were"). It is also potentially present in the form , which appears on a broken inscription in the Vestinian language. In Classical Latin, imperfect forms of this type for the first and second conjugations are analyzable as the verb stem plus the infix, as in and . In contrast, third conjugation forms such as and some fourth conjugation forms such as contain an vowel that is not otherwise apart of the stem. However, there is attestation of fourth conjugation imperfect forms such as in the works of Old Latin authors such as Plautus and Terence, who utilized this unique type of imperfect instead of the standard Classical forms. According to the philologist Wolfgang de Melo, it is likely that Classical forms such as were a later innovation on the model of second conjugation forms such as after the suffix had been reanalyzed as instead of .
The new suffix was itself perhaps from the combination of the root ("to be") and the same -extension found in the imperfect forms of the verb (i.e. ). If a connection between the imperfective suffix and the root is accepted, then a connection with becomes problematic. This stem of this Oscan verb is also typically derived from , in which case, it would perhaps be unlikely for the term to be modified with a suffix derived from the same root. It is often presumed that the earlier form of this suffix took the shape , itself possibly continued by rare Latin forms such as and negative imperatives such as , though this connection requires a reduction to in enclitic position. The exact Indo-European origin of the form is unclear; it is sometimes considered a reflex of an aorist belonging to a root , which is perhaps a regularized form of the root . Sihler is, however, critical of the aorist explanation for this formation, arguing that aorists otherwise develop into Latin perfects and it would be unusual for a perfective formation denoting completed action to evolve into an imperfective term describing continuous action.
The form perhaps appeared in set phrases wherein in functioned alongside other verbs, with which it may have eventually merged. Similar processes probably underlie the future forms in various Romance languages: Proto-Romance forms such as , for instance, yielded French future forms such as . If such a development occurred in Proto-Italic, according to Sihler, the exact series of necessary sound changes and analogical remodeling required would be both complex and unique to the particular phrases featuring , thereby inhibiting any attempt to reconstruct the pre-history of this formation. One theory suggests that the particular set phrase in question originally consisted of an infinitive term alongside a form of (e.g. ). However, sentences containing an infinitive combined with any form of ("to be") are otherwise entirely unknown in Latin, excluding impersonal constructions (i.e. "," "one may know"). It has also been suggested that the morpheme was affixed to the bare verbal root (e.g. - + ), though this theory is itself marred by the lack of any parallel instances wherein the verbal root functioned as an independent entity in Latin. Alternatively, the linguist Phillip Baldi has suggested that the sequence perhaps shifted into in rapid speech, thereby allowing for a phrase such as to develop into , whereafter the would be deleted prior to a voiced stop, ultimately yielding the attested form . In further support of this theory, Baldi notes that participles can freely mix with inflected forms of the verb in Latin.
Independently, within the Latino-Faliscan languages, a new future tense marked by the same infix was created on the basis of the imperfect (e.g. , ; , ). The new future displaced the inherited future forms for the first and second conjugations sometime in Latin prehistory, whereas it saw much more limited success in the third and fourth conjugations. No b-futures survive for the third conjugation, whereas there are some attested fourth conjugation futures in Old Latin writers such as Plautus and Terence, who use forms such as and , though such terms are still uncommon.
Subjunctive
Present subjunctive During the transition from Proto-Indo-European to Latin and Sabellic, the subjunctive and optative were merged, resulting in the general Italic subjunctive. However, the linguist Frederick Kortlandt disputes whether the optative had been lost by the time of Proto-Italic, noting that the continuation of optative forms may be attested in terms such as Latin , which perhaps continues the Proto-Indo-European athematic optative infix . This same athematic infix is probably also reflected in the subjunctive forms of the irregular verbs ("to be"), ("to eat"), and ("to want") ( > Old Latin ). It is possible that the original Proto-Indo-European optative forms the basis for the mood known in Latin as the subjunctive, whereas the original Proto-Indo-European subjunctive paradigm gave rise to the standard Latin future. If the PIE subjunctive is reconstructed with alternation between the and vowels in the inflectional endings, then the -vowel must have been generalized throughout the entire paradigm in Latin, thus yielding inflectional endings such as (e.g. ) in contrast to Ancient Greek (";" e.g. , "").
The Latin first conjugation marks its subjunctive endings utilizing the morpheme , such as in , a subjunctive form of . It is possible that this paradigm derives directly from the Proto-Indo-European thematic subjunctive of verbs marked by the suffix or from the athematic optative morpheme . According to the linguist Jay Jasanoff, a proponent of the latter hypothesis, the Latin -subjunctive spread from the small class of verbs that originated from athematic -factitives. In this scenario, a Latin subjunctive such as (subjunctive of , "to make new") would derive from Proto-Italic , itself from an athematic optative . Regardless, in either case, the inevitable Proto-Italic first conjugation sequence could have perhaps have, following the loss of intervocalic yod, contracted into Latin . Such a contraction would be unusual for Latin, as the expected outcome of a sequence would be , which would render the present subjunctive and indicative paradigms identical. To resolve this issue, Rix presumes that the irregular sound development served to prevent the present indicative and subjunctive from becoming indistinguishable. Alternatively, Sihler posits that the sequence metathesized to under the influence of optative forms such as . The proposed contraction did not occur in the Sabellic branch, thereby permitting first conjugation subjunctive formations such as Umbrian , the ending of which derives from earlier .
All other Latin conjugations form their subjunctive paradigms utilizing the morpheme , such as in Latin , a subjunctive form of . Likewise, terms such as Umbrian attest to the existence of a similar formation in Sabellic. Outside of Italic, the a-subjunctive may be paralleled in Old Irish—compare Old Irish subjunctive forms such as to the Latin equivalent . In Old Irish, the a-subjunctive is formed from the verbal root itself, which is paralleled by a select few archaic formations in Italic, such as Hernican and Old Latin . Traditionally, the a-subjunctive has been interpreted as an Italo-Celtic development which possibly ousted the original Proto-Indo-European thematic optative. However, this theory is controversial as it provides no sufficient explanation for the independent development of the -morpheme in Italo-Celtic. Moreover, Kortlandt suggests that the Old Irish and Latin a-subjunctive forms cannot be derived from a single paradigm. For instance, Kortlandt notes that the hypothetical pre-forms , though capable of producing Latin , , and , would output Old Irish , , and , instead of the attested forms , , and .