Development of nominal endings from PIE to Latin

Note "(<)" abbreviates "(fused with last stem sound (consonant or vowel))", "(<<)" abbreviates "(fused with last stem vowel + consonant)", "(–)" abbreviates "(previous vowel is lengthened)", "(v)" abbreviates "(previous vowel is shortened)", "?(x)" means "possibly x", "(x|y)" means "either x or y".

Late Proto-Indo-European (simplified)

This table omits the dual number, of which the development of Latin retains few if any traces. Not depicted, but part of Italic development: Pre-laryngeal e-ablauts were already laryngeally colored (i.e. phonetically eh₂ has become ah₂, and h₂e h₂a; Fortson p. 63; Meier-Brügger p. 244; Meiser p. 29; Sihler p. 42-50). Variations in athematic root/stem accent/ablaut are mostly ignored except for a few cases where stem ablauts easily fuse with the endings proper (i.e. i- and u-stems; see "genus"/"generis" for an ablaut graduation continued straight into Latin). Regarding consonant-stem singular nominatives, most scholars assume an -s ending, and for that -s ending to be dropped (with compensatory lengthening of the preceding vowel) after non-stops. Beekes p. 186 though assumes the opposite, consonant-stem nouns originally ending in -∅, with -s only copied over later from the o-stem nouns. For stems ending in -n- and -(d|t)-, the stem-final consonant may have dropped too, with any subsequent -s only kept in the latter (example *kwō from *kwon-: Beekes p. 194, Sihler p. 100, 230; example *pōs from *pod-: Sihler p. 130, 280; not depicted: Sihler p. 280 assumes for *pod- that the dental stop was later analogically restored, with the lengthened vowel kept: *pōts .) The accusative plural endings in -ns probably result from suffixing the plural nominative -s to the accusative singular -m (Klein/Joseph/Fritz p. 2084; Meier-Brügger p. 295). Not depicted: a thematic singular genitive in -os, only to be found in the Anatolian branch (Klein/Joseph/Fritz p. 2086), with -os?(i)o probably an early import from the pronouns (Beekes p. 213); an alternate athematic singular locative without ending (Sihler p. 248, 253); the u-stem dative singular ending in -i rather than -ei according to Beekes (p. 203), without explanation contrary to all other opinions. An athematic singular genitive/ablative in mere zero-grade ending -s may have been productiv in Italic for the i- and u-stems (with e/o-ablaut grade of the suffix), and is therefore depicted as a possibility, while a possible sibling for the consonantal stems is not depicted, since it only shows traces in other branches (Klein/Joseph/Fritz p. 2086).

thematic cons. -i- -u-
-(g|k|b|p)- -(d|t)- -n- &c.
sg:N -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s
sg:V -e -∅ -ei (<) -(e|o)u (<)
sg:Ac -om -m
sg:n -om -∅
sg:G -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -(e|o)us (<), ?(-(e|o)s)
sg:Ab -oHed -(e|o)s -(e|o)is (<), ?(-(e|o)s) -(e|o)us (<), ?(-(e|o)s)
sg:D -ōi -ei -eiei (<) -(e|o)uei (<)
sg:L -(e|o)i -i -ēi (<) -(ē|ō)u (<)
sg:I -(e|o)h₁ -?(e)h₁, -(bʰ|m)i -?(e)h₁, -(bʰ|m)i
pl:N/V -ōs -es -eies (<) -(e|o)ues (<)
pl:Ac -ons -ns
pl:n -eh₂ -h₂
pl:G -(o|oHo|ō)m -(o|oHo|ō)m
pl:D/Ab -o?(i)(bʰ?(i)|m)os -(bʰ?(i)|m)os
pl:L -oisu -su
pl:I -ōis -(bʰ|m)i?(s)

Late Proto-Indo-European eh₂ (feminine) declension

Possibly after the split-off of the Anatolian branch, an explicitely feminine declension emerges, marked by a -eh₂- stem suffix to which athematic endings are appended (Beekes p. 189; Fortson p. 132-133; Meier-Brügger p. 323-324, 334-335, 421; Meiser p. 131; Sihler p. 245, 266). The singular vocative is a special case, with different theories about its form and the logic behind it.

eh₂ thematic cons. -i- -u-
-(g|k|b|p)- -(d|t)- -n- &c.
sg:N -eh₂ -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s
sg:V -(e?(h₂)|h₂) -e -∅ -ei (<) -(e|o)u (<)
sg:Ac -eh₂m -om -m
sg:n -om -∅
sg:G -eh₂(e|o)s -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -(e|o)us (<), ?(-(e|o)s)
sg:Ab -eh₂(e|o)s -ōd -(e|o)s -(e|o)is (<), ?(-(e|o)s) -(e|o)us (<), ?(-(e|o)s)
sg:D -eh₂ei -ōi -ei -eiei (<) -(e|o)uei (<)
sg:L -eh₂i -(e|o)i -i -ēi (<) -(ē|ō)u (<)
sg:I -eh₂eh₁, -eh₂(bʰ|m)i -(e|o)h₁ -?(e)h₁, -(bʰ|m)i -?(e)h₁, -(bʰ|m)i
pl:N/V -eh₂es -ōs -es -eies (<) -(e|o)ues (<)
pl:Ac -eh₂ns -ons -ns
pl:n -eh₂ -h₂
pl:G -eh₂(o|oHo|ō)m -(o|oHo|ō)m -(o|oHo|ō)m
pl:D/Ab -eh₂(bʰ?(i)|m)os -o?(i)(bʰ?(i)|m)os -(bʰ?(i)|m)os
pl:L -eh₂su -oisu -su
pl:I -eh₂(bʰ|m)is -ōis -(bʰ|m)i?(s)

Late Proto-Indo-European bʰ/m-branching

In the Italic branch (among others), the endings starting in either -m- or -bʰ- generalize the latter.

eh₂ thematic cons. -i- -u-
-(g|k|b|p)- -(d|t)- -n- &c.
sg:N -eh₂ -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s
sg:V -(e?(h₂)|h₂) -e -∅ -ei (<) -(e|o)u (<)
sg:Ac -eh₂m -om -m
sg:n -om -∅
sg:G -eh₂(e|o)s -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -(e|o)us (<), ?(-(e|o)s)
sg:Ab -eh₂(e|o)s -ōd -(e|o)s -(e|o)is (<), ?(-(e|o)s) -(e|o)us (<), ?(-(e|o)s)
sg:D -eh₂ei -ōi -ei -eiei (<) -(e|o)uei (<)
sg:L -eh₂i -(e|o)i -i -ēi (<) -(ē|ō)u (<)
sg:I -eh₂eh₁, -eh₂bʰi -(e|o)h₁ -?(e)h₁, -bʰi -?(e)h₁, -bʰi
pl:N/V -eh₂es -ōs -es -eies (<) -(e|o)ues (<)
pl:Ac -eh₂ns -ons -ns
pl:n -eh₂ -h₂
pl:G -eh₂(o|oHo|ō)m -(o|oHo|ō)m -(o|oHo|ō)m
pl:D/Ab -eh₂bʰ?(i)os -o?(i)bʰ?(i)os -bʰ?(i)os
pl:L -eh₂su -oisu -su
pl:I -eh₂bʰis -ōis -bʰi?(s)

Proto-Italic loss of laryngeals, and Proto-Indo-European laryngeal coloring

The laryngeals disappear. Vowels neighboring lost laryngeals are left in their phonetic laryngeal coloring (i.e. -e- as -a- when neighbor to -h₂-), and lengthened if the lost laryngeal came after and was word-final or followed by a consonant (Beekes p. 149-150; Clackson/Horrocks p. 9-10; Fortson p. 63-64; Meier-Brügger p. 244; Meiser p. 28-30, 56-57; Sihler p. 42-50). Word-final post-consonantal laryngeals (maybe even after -u: Meiser p. 147) become -a (Sihler p. 99; Meiser p. 107). From these changes, vowel neighborships -aa- (a-stem dative singular), -āā- (a-stem instrumental singular) and (possibly with the a-stem genitive/ablative singular) -ao- emerge that contract to -ā- (Meiser p. 87-88; Sihler p. 83). Post-laryngeal continuation of former eh₂-stem singular instrumental ending -eh₂bʰi is not directly discussed in the literature, but would have given -ābʰi by these rules. The a-stem singular vocative form reflects no single rule path, but an uncertainty that fits the variety of theories about its original form. The i/u-stem singular instrumentals in -ī (<) and -ū (<) are not directly attested for Italic, but are the logical result of the above rules also found in Vedic and Avestan and speculatively included here as such (Fortson p. 116, Sihler p. 312, 323).

a o cons. -i- -u-
-(g|k|b|p)- -(d|t)- -n- &c.
sg:N -ā -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s
sg:V -(a|ā) -e -∅ -ei (<) -(e|o)u (<)
sg:Ac -ām -om -m
sg:n -om -∅
sg:G -ās -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -(e|o)us (<), ?(-(e|o)s)
sg:Ab -ās -ōd -(e|o)s -(e|o)is (<), ?(-(e|o)s) -(e|o)us (<), ?(-(e|o)s)
sg:D -āi -ōi -ei -eiei (<) -(e|o)uei (<)
sg:L -(a|ā)i -(e|o)i -i -ēi (<) -(ē|ō)u (<)
sg:I -ā -(ē|ō) -(ē|a), -bʰi -ī (<), -bʰi -ū (<), -bʰi
pl:N/V -ās -ōs -es -eies (<) -(e|o)ues (<)
pl:Ac -āns -ons -ns
pl:n -ā -a -ī (<) -(ua|ū) (<)
pl:G -a(o|ō)m -(o|ō)m -(o|ō)m
pl:D/Ab -ābʰ?(i)os -o?(i)bʰ?(i)os -bʰ?(i)os
pl:L -āsu -oisu -su
pl:I -ābʰis -ōis -bʰi?(s)

Proto-Italic fusion of (dative-)ablative and instrumental

Ablative and instrumental fuse to the new ablative (Clackson/Horrocks p. 14-15; Fortson p. 281; Meiser p. 128; Sihler p. 252-253, 258), in the plural together with the formally already identical dative (Meiser p. 128-129, Sihler p. 264): The o-stem instrumental plural ending -ōis replaces the dative/ablative ending. The a-stem dative/ablative plural ending is re-shaped after the result to -āis. In the singular, the pattern of the o-stem ablative ending -ōd is generalized to most other stems. (There may have been an intermediate stage where the singular instrumental endings -ā, -ī, -ū replaced the old ablative endings not as distinctive as -ōd, and were then re-shaped by mere addition of its -d.) Whether this pattern extended to the consonantal stems as -ēd is questionable (Weiss p. 202, 238 vs. Meiser p. 138; Sihler p. 285 tentatively proposes the possibility of a continuation in mere instrumental -ē without final -d).

a o cons. -i- -u-
-(g|k|b|p)- -(d|t)- -n- &c.
sg:N -ā -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s
sg:V -(a|ā) -e -∅ -ei (<) -(e|o)u (<)
sg:Ac -ām -om -m
sg:n -om -∅
sg:G -ās -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -(e|o)us (<), ?(-(e|o)s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<)
sg:D -āi -ōi -ei -eiei (<) -(e|o)uei (<)
sg:L -(a|ā)i -(e|o)i -i -ēi (<) -(ē|ō)u (<)
pl:N/V -ās -ōs -es -eies (<) -(e|o)ues (<)
pl:Ac -āns -ons -ns
pl:n -ā -a -ī (<) -(ua|ū) (<)
pl:G -a(o|ō)m -(o|ō)m -(o|ō)m
pl:D/Ab -āis -ōis -bʰ?(i)os
pl:L -āsu -oisu -su

Proto-Italic influence of Osthoff's Law

As per Osthoff's Law, long vowels before resonants are shortened if the latter are followed by consonants (Meiser p. 75; Sihler p. 58).

a o cons. -i- -u-
-(g|k|b|p)- -(d|t)- -n- &c.
sg:N -ā -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s
sg:V -(a|ā) -e -∅ -ei (<) -(e|o)u (<)
sg:Ac -ām -om -m
sg:n -om -∅
sg:G -ās -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -(e|o)us (<), ?(-(e|o)s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<)
sg:D -āi -ōi -ei -eiei (<) -(e|o)uei (<)
sg:L -(a|ā)i -(e|o)i -i -ēi (<) -(ē|ō)u (<)
pl:N/V -ās -ōs -es -eies (<) -(e|o)ues (<)
pl:Ac -ans -ons -ns
pl:n -ā -a -ī (<) -(ua|ū) (<)
pl:G -a(o|ō)m -(o|ō)m -(o|ō)m
pl:D/Ab -ais -ois -bʰ?(i)os
pl:L -āsu -oisu -su

Proto-Italic (sound) changes to the plural athematic ablative/dative

The voiced aspirate -bʰ- in medial positions transforms to the voiceless labiodental fricative -f- (Sihler p. 139-141) or the voiced bilabial fricative -β- (Meiser p. 101-103; Clackson/Horrocks p. 51-52; Weiss p. 149-150; Klein/Joseph/Fritz p. 744-745). An -i- that follows this sound in some Indo-European branches disappears in Proto-Italic (Klein/Joseph/Fritz p. 2088).

a o cons. -i- -u-
-(g|k|b|p)- -(d|t)- -n- &c.
sg:N -ā -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s
sg:V -(a|ā) -e -∅ -ei (<) -(e|o)u (<)
sg:Ac -ām -om -m
sg:n -om -∅
sg:G -ās -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -(e|o)us (<), ?(-(e|o)s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<)
sg:D -āi -ōi -ei -eiei (<) -(e|o)uei (<)
sg:L -(a|ā)i -(e|o)i -i -ēi (<) -(ē|ō)u (<)
pl:N/V -ās -ōs -es -eies (<) -(e|o)ues (<)
pl:Ac -ans -ons -ns
pl:n -ā -a -ī (<) -(ua|ū) (<)
pl:G -a(o|ō)m -(o|ō)m -(o|ō)m
pl:D/Ab -ais -ois -(f|β)os
pl:L -āsu -oisu -su

Proto-Italic vocative losses/merges/contractions

In the athematic stems, the singular vocative disappears/merges with the nominative (Sihler p. 283, 316, 322; Meiser p. 138-139, 147). Latin "Iuppiter" preserves a trace of the endingless u-stem singular vocative, as composite of vocative *dieu with *pāter (Meiser p. 138, 144; Weiss p. 200). In the io-stems, the vocative -i-e may possibly contract to -ī (safely attested in Old Latin, but a bit controversial as a Proto-Italic stage: Weiss p. 221; Leumann p. 424-425; Sihler p. 257-258).

a o cons. -i- -u-
&c. -i- -(g|k|b|p)- -(d|t)- -n- &c.
sg:N -ā -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s
sg:V -(a|ā) -e (-e)|(-ī (<))
sg:Ac -ām -om -m
sg:n -om -∅
sg:G -ās -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -(e|o)us (<), ?(-(e|o)s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<)
sg:D -āi -ōi -ei -eiei (<) -(e|o)uei (<)
sg:L -(a|ā)i -(e|o)i -i -ēi (<) -(ē|ō)u (<)
pl:N/V -ās -ōs -es -eies (<) -(e|o)ues (<)
pl:Ac -ans -ons -ns
pl:n -ā -a -ī (<) -(ua|ū) (<)
pl:G -a(o|ō)m -(o|ō)m -(o|ō)m
pl:D/Ab -ais -ois -(f|β)os
pl:L -āsu -oisu -su

Proto-Italic assimilation of e to o before u

Before u, (short) e becomes o (Sihler p. 40, 56; Meiser p. 59; Klein/Joseph/Fritz p. 746). The u-stem plural nominative/locative in -oues is here depicted as the regular result, although Sihler p. 325 states: "for some reason, that is not generally regarded as likely."

a o cons. -i- -u-
&c. -i- -(g|k|b|p)- -(d|t)- -n- &c.
sg:N -ā -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s
sg:V -(a|ā) -e (-e)|(-ī (<))
sg:Ac -ām -om -m
sg:n -om -∅
sg:G -ās -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -ous (<), ?(-(e|o)s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<)
sg:D -āi -ōi -ei -eiei (<) -ouei (<)
sg:L -(a|ā)i -(e|o)i -i -ēi (<) -(ē|ō)u (<)
pl:N/V -ās -ōs -es -eies (<) -oues (<)
pl:Ac -ans -ons -ns
pl:n -ā -a -ī (<) -(ua|ū) (<)
pl:G -a(o|ō)m -(o|ō)m -(o|ō)m
pl:D/Ab -ais -ois -(f|β)os
pl:L -āsu -oisu -su

Proto-Italic contraction of -eie- to -ē-, and -ēi to -ei-

The singular i-stem dative ending in -eiei is contracted to -ei, either by haplology (Meiser p. 139, Szemerényi p. 187), or by a regular Proto-Italic loss of intervocalic -i- (Sihler p. 188; Meiser p. 91; Klein/Joseph/Fritz p. 747) shortening -eiei to -eei, which regularly becomes -ēi (Sihler p. 83, Meiser p. 87), then -ei. Loss of intervocalic -i- would also explain the development of the i-stem nominative/vocative plural (Clackson/Horrocks p. 102, Szemerényi p. 187).

a o cons. -i- -u-
&c. -i- -(g|k|b|p)- -(d|t)- -n- &c.
sg:N -ā -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s
sg:V -(a|ā) -e (-e)|(-ī (<))
sg:Ac -ām -om -m
sg:n -om -∅
sg:G -ās -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -ous (<), ?(-(e|o)s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<)
sg:D -āi -ōi -ei -ei (<) -ouei (<)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -ās -ōs -es -ēs (<) -oues (<)
pl:Ac -ans -ons -ns
pl:n -ā -a -ī (<) -(ua|ū) (<)
pl:G -a(o|ō)m -(o|ō)m -(o|ō)m
pl:D/Ab -ais -ois -(f|β)os
pl:L -āsu -oisu -su

Proto-Italic plural ablative/dative/locative

In the plurals, dative/ablative and locative fuse to one common form per stem. In the a/o-stems, this form is identical to the plural dative/ablative's original form. In the o-stems, the result of -ois would have been equal to the locative ending -oisu on loss of its -u (Meiser p. 74), which may have happened at the same time. This formal identity between dative/ablative and locative may have spread to the other stems. The extension of the athematic dative/ablative to the locative is somewhat speculative (Meiser p. 129 "mag verwendet worden sein", offers a Sabellic example though similar to the later Latin outcome).

a o cons. -i- -u-
&c. -i- -(g|k|b|p)- -(d|t)- -n- &c.
sg:N -ā -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s
sg:V -(a|ā) -e (-e)|(-ī (<))
sg:Ac -ām -om -m
sg:n -om -∅
sg:G -ās -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -ous (<), ?(-(e|o)s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<)
sg:D -āi -ōi -ei -ei (<) -ouei (<)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -ās -ōs -es -ēs (<) -oues (<)
pl:Ac -ans -ons -ns
pl:n -ā -a -ī (<) -(ua|ū) (<)
pl:G -a(o|ō)m -(o|ō)m -(o|ō)m
pl:D/Ab/L -ais -ois -(f|β)os

Proto-Italic a-stem plural genitive

The a-stem plural genitive ending is re-shaped after the respective pronoun form.

a o cons. -i- -u-
&c. -i- -(g|k|b|p)- -(d|t)- -n- &c.
sg:N -ā -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s
sg:V -(a|ā) -e (-e)|(-ī (<))
sg:Ac -ām -om -m
sg:n -om -∅
sg:G -ās -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -ous (<), ?(-(e|o)s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<)
sg:D -āi -ōi -ei -ei (<) -ouei (<)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -ās -ōs -es -ēs (<) -oues (<)
pl:Ac -ans -ons -ns
pl:n -ā -a -ī (<) -(ua|ū) (<)
pl:G -ās(o|ō)m -(o|ō)m -(o|ō)m
pl:D/Ab/L -ais -ois -(f|β)os

Proto-Italic u-stem singular dative

A u-stem singular dative in -ou appears, either in analogy to the i-stem ending (-ei to -ou: Sihler p. 323; Weiss p. 251; Bakkum p. 146; Meiser p. 146; Klein/Joseph/Fritz p. 761), or derived from the locative (Sihler). (Note that Meiser and Bakkum seem to see this ending as older than -ouei or -uei, with the latter modelled later after the consonant-stem endings.)

a o cons. -i- -u-
&c. -i- -(g|k|b|p)- -(d|t)- -n- &c.
sg:N -ā -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s
sg:V -(a|ā) -e (-e)|(-ī (<))
sg:Ac -ām -om -m
sg:n -om -∅
sg:G -ās -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -ous (<), ?(-(e|o)s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -ās -ōs -es -ēs (<) -oues (<)
pl:Ac -ans -ons -ns
pl:n -ā -a -ī (<) -(ua|ū) (<)
pl:G -ās(o|ō)m -(o|ō)m -(o|ō)m
pl:D/Ab/L -ais -ois -(f|β)os

Proto-Italic or later i/u-stem neuters

The i- and u-stem neuter plural ending is re-shaped analogically to the either the o-stem (Sihler p. 317, 326) or the consonantal stem (Meiser p. 144, p. 146-147) neuter plural ending; alternatively, according to Meiser (p. 147), the u-stem might have just continued an original -ua ending. The plural change seems (Proto-)Italic (Meiser p. 144, Weiss p. 243, 252, Sihler p. 317, 326), less clear is the distribution beyond Latin of a (slightly controversial, see Meiser) u-stem singular in -ū, which may have been influenced by the original plural neuter before that got replaced in this step (Meiser p. 146; Klein/Joseph/Fritz p. 760; Weiss p. 252; Sihler p. 323).

a o cons. -i- -u-
&c. -i- -(g|k|b|p)- -(d|t)- -n- &c.
sg:N -ā -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s
sg:V -(a|ā) -e (-e)|(-ī (<))
sg:Ac -ām -om -m
sg:n -om -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -ous (<), ?(-(e|o)s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -ās -ōs -es -ēs (<) -oues (<)
pl:Ac -ans -ons -ns
pl:n -ā -a -(a|ā)
pl:G -ās(o|ō)m -(o|ō)m -(o|ō)m
pl:D/Ab/L -ais -ois -(f|β)os

Emergence of the ē-declension between Proto-Italic and Early Latin

The declension of some words with stems in -eu- and -ei- merges into its own dedicated declension, with stems ending in -ē-. The clearest influence is Proto-Indo-European *dieu-, which by Stang's Law had a singular PIE accusative *diē-m (rather than *dieu-m); this diē- was generalized as a stem to the singular nominative and beyond (Meiser p. 143-144; Weiss p. 253-254; Sihler p. 337-339). Similarly, i-stem *reHi-, after laryngeal loss *rei- or *rēi-, would, through loss of intervocalic -i- before endings starting in a vocal, produce a new stem rē- (Meiser p. 148; Weiss p. 254; Sihler p. 340-341). Some Italic forms other than Latin would make sense as parts of an ē-declension, but the evidence there is far from abundant or clear (Sihler p. 342; Bakkum p. 147; Klein/Joseph/Fritz p. 761, 763), so the declension's emergence should lie somwehere between Proto-Italic and first epigraphic attestations of its peculiarities in Latino-Faliscan around the 3rd century BCE. The earliest stages attested or assumed keep the regular endings of the athematic declensions, though possibly contracted (singular dative and genitive, plural nominative), or show analogy to a-declension endings with -ā- replaced by the stem's -ē-. Plurals are barely attested beyond rēs and diēs, and except for the nominative explained (Sihler p. 343) as analogies to other declensions, perfectly so by much later stages of those than what is highly speculatively depicted here. TODO: depict precursors in i- and u-stems, and thus the history of "bovis" too.

a o cons. -i- -u- -ē-
&c. -i- -(g|k|b|p)- -(d|t)- -n- &c.
sg:N -ā -os -?s (-?s)|(-s (<,–))|(-∅ (–)) -?n (<,–) -∅ (–) -s -s
sg:V -(a|ā) -e (-e)|(-ī (<))
sg:Ac -ām -om -m -m
sg:n -om -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -ous (<), ?(-(e|o)s) -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -ās -ōs -es -ēs (<) -oues (<) -s
pl:Ac -ans -ons -ns ?(-ns)
pl:n -ā -a -(a|ā)
pl:G -ās(o|ō)m -(o|ō)m -(o|ō)m ?(-s(o|ō)m)
pl:D/Ab/L -ais -ois -(f|β)os ?(-(f|β)os)

Athematic nominative singulars between Proto-Italic and Early Latin

Status of the nominative singular in athematic stems somewhere between Proto-Italic and Early Latin: For stems in -t-, -g-, -k- ("c" in Latin), -gʷʰ-,-gʰ-, -kʷ-, -b-, -p-, -m-, all further development now safely points to an ending -s. For stems in -ti-, the nominative singular in -ti-s is often syncopated (probably after two mores – Meiser p. 73, Leumann p. 98) to -t-s (at first). Forms in -(d|t)-s like the aforementioned assimilate further to -s-s. If a stem's -ti- is preceded by -n-, the preceding vowel is probably already lengthened (similarly with -di-, as in frōns/frondis – Sihler p. 316). Syncopated and unsyncopated forms of ti-stem nominative singular endings co-exist at least until Ennius and Plautus (Leumann p. 449). A stem ending in a velar fuses with the nominative ending -s to what's written as "-x" in Latin; notably, -gʰ and -gʷʰ- become -h- and -u- word-internally (Sihler p. 156, 158; Weiss p. 77, 79; Meiser p. 104), so the new stems ending in these also sport a nominative in -x. For stems ending -b-, the -b-s result was sometimes written as -p-s, and probably always spoken that way. For the m-stem we have (only?) "hiems", which was sometimes written "hiemps". A dental stop after a cosonant is eliminated in word-final position ("lac"/"lactis", "cor"/"cordis"). Latin shows some neuter ll-stems (genitives "mellis", "fellis" with the latter originally a stem in -ln, Sihler p. 298), where the ending simplifies -ll to -l. For n-stems, any (Pre?-)PIE *-o-n-s has safely become -ō (for "carō"/"carnis", the genitive can be explained by post-r-vowel syncopation of what otherwise could have been "carinis": Sihler p. 295), but some endings in -en also emerge, possibly from gender change or compounds with weakened root "can-" (Sihler p. 296).

a o cons. -i- -u- -ē-
&c. -i- -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -s- -n(d|t)i- -ti- &c
sg:N -ā -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–) -∅ (–) -ss ?(–) (<<), -s -ss (<<), -s -s -s -s
sg:V -(a|ā) -e (-e)|(-ī (<))
sg:Ac -ām -om -m -m
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -∅ -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -os?(i)o -(e|o)s -(e|o)is (<), ?(-(e|o)s) -ous (<), ?(-(e|o)s) -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -ās -ōs -es -ēs (<) -oues (<) -s
pl:Ac -ans -ons -ns ?(-ns)
pl:n -ā -a -(a|ā)
pl:G -ās(o|ō)m -(o|ō)m -(o|ō)m ?(-s(o|ō)m)
pl:D/Ab/L -ais -ois -(f|β)os ?(-(f|β)os)

The o-stem singular genitive from possibly Celtic to Latino-Faliscan

Proto-Indo-European -os?(i)o is continued as -osio possibly even into Archaic Latin (Bakkum p. 132-134; Meiser p. 133-145; Weiss p. 222). An ending -ī of uncertain origin replaces this in Latin, but also turns up in Venetic (which, if part of the Italic family, split off from it early; Beekes p. 25; Fortson p. 127, 465-466), Faliscan, and even Celtic languages. This may continue a Proto-Italic or even Proto-Italo-Celtic ending. Or it may be the result of more recent morphological borrowing between languages in close contact on the Italian Peninsula (Clackson/Horrocks p. 32). The -ī ending may continue a suffix -iH that marks familial affiliation (Beekes p. 212; Meier-Brügger p. 333; Meiser p. 135). After -i-, the new ending contracted with the preceding vowel.

a o cons. -i- -u- -ē-
&c. -i- -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -m- -l- -n- -r- -s- -n(d|t)i- -ti- &c
sg:N -ā -os -s -(b|p)s (<) -x (<) -ss (<) -?(p)s -∅ (–) -ō (<<, < for stem in -rn-), -en (<<) -∅ (–) -∅ (–) -ss ?(–) (<<), -s -ss (<<), -s -s -s -s
sg:V -(a|ā) -e (-e)|(-ī (<))
sg:Ac -ām -om -m -m
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -∅ -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -osio, -ī -osio, -ī (<) -(e|o)s -(e|o)is (<), ?(-(e|o)s) -ous (<), ?(-(e|o)s) -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -ās -ōs -es -ēs (<) -oues (<) -s
pl:Ac -ans -ons -ns ?(-ns)
pl:n -ā -a -(a|ā)
pl:G -ās(o|ō)m -(o|ō)m -(o|ō)m ?(-s(o|ō)m)
pl:D/Ab/L -ais -ois -(f|β)os ?(-(f|β)os)

Proto-Italic or Proto-Latino-Faliscan vowel shortening before -m

Long vowels before -m are shortened (Sihler p. 78, Meiser p. 77). The chronology is disputed. The shortening is shared with Celtic and Slavic, so might be early; but regular sound change would then have turned endings in -am into -em (Sihler p. 65). There is also evidence for -ām in Italic Oscan. Meiser (as do Weiss p. 233; Klein/Joseph/Fritz p. 755-757) opts for a late change, while Sihler explains away outliers to the expected development from an early change by leveling / influence from other endings that contain -a- (in Latin) or -ā- (in Oscan). Weiss p. 224 proposes a later shortening happening independently in Sabellic.

a o cons. -i- -u- -ē-
&c -i- -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -s- -n(d|t)i- -ti- &c
sg:N -ā -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–) -∅ (–) -ss ?(–) (<<), -s -ss (<<), -s -s -s -s
sg:V -(a|ā) -e (-e)|(-ī (<))
sg:Ac -am -om -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -∅ -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -osio, -ī -osio, -ī (<) -(e|o)s -(e|o)is (<), ?(-(e|o)s) -ous (<), ?(-(e|o)s) -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -ās -ōs -es -ēs (<) -oues (<) -s
pl:Ac -ans -ons -ns ?(-ns)
pl:n -ā -a -(a|ā)
pl:G -āsom -om -om ?(-som)
pl:D/Ab/L -ais -ois -(f|β)os ?(-(f|β)os)

Proto-Latino-Faliscan thematic plural animate nominatives/vocatives

The o-stem plural animate nominative is replaced by the respective pronoun ending (contrast with Sabellic, which went the opposite way: Sihler p. 261). The a-stem plural animate nominative is re-modelled after this influence, although the -ās form is still used rarely in later centuries. Sihler p. 271 calls these usages "inherited", but they only seem to show up relatively late and might therefore be (Weiss p. 235) Sabellisms or even mere confusions with the accusative.

a o cons. -i- -u- -ē-
&c -i- -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -s- -n(d|t)i- -ti- &c
sg:N -ā -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–) -∅ (–) -ss ?(–) (<<), -s -ss (<<), -s -s -s -s
sg:V -(a|ā) -e (-e)|(-ī (<))
sg:Ac -am -om -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -∅ -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -osio, -ī -osio, -ī (<) -(e|o)s -(e|o)is (<), ?(-(e|o)s) -ous (<), ?(-(e|o)s) -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -(a|ā)i, ?(-ās) -oi -es -ēs (<) -oues (<) -s
pl:Ac -ans -ons -ns ?(-ns)
pl:n -ā -a -(a|ā)
pl:G -āsom -om -om ?(-som)
pl:D/Ab/L -ais -ois -(f|β)os ?(-(f|β)os)

Proto-Latino-Faliscan treatment of syllabic nasals

Syllabic nasals turn into nasals prefixed with an -e- (in Sabellic maybe with an -a-: Meiser p. 65; Fortson p. 277; Sihler p. 96; Weiss p. 94-95). In athematic n-stem neuters, the suffix was a syllabic nasal in the strong cases, while the weak cases were ablauting (to whatever later became -in- in Latin), so this accounts for the singular nominatives in -en there (Sihler p. 297-298).

a o cons. -i- -u- -ē-
&c -i- -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -s- -n(d|t)i- -ti- &c
sg:N -ā -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–) -∅ (–) -ss ?(–) (<<), -s -ss (<<), -s -s -s -s
sg:V -(a|ā) -e (-e)|(-ī (<))
sg:Ac -am -om -em -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -osio, -ī -osio, -ī (<) -(e|o)s -(e|o)is (<), ?(-(e|o)s) -ous (<), ?(-(e|o)s) -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -(a|ā)i, ?(-ās) -oi -es -ēs (<) -oues (<) -s
pl:Ac -ans -ons -ens -ns ?(-ns)
pl:n -ā -a -(a|ā)
pl:G -āsom -om -om ?(-som)
pl:D/Ab/L -ais -ois -(f|β)os ?(-(f|β)os)

Proto-Latino-Faliscan and Early Latin i- and u-stem singular genitives

For both i- and u-stems, a singular genitive ending is found that may either continue a possible Proto-Indo-European -(e|o)s ending for these very same stems that would have existed alongside the -(e|o)is and -(e|o)us endings, or import the identical consonant-stem ending. In the u-stems, this ending is merely appended to the suffix. In the i-stems, the e-grade may either have replaced the stem-final -i- to -e-s, or (Sihler p. 316-317; questioned by Weiss p. 245) syncopated from an intermediary -i-(e|o)s to -i-s.

a o cons. -i- -u- -ē-
&c -i- -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -s- -n(d|t)i- -ti- &c
sg:N -ā -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–) -∅ (–) -ss ?(–) (<<), -s -ss (<<), -s -s -s -s
sg:V -(a|ā) -e (-e)|(-ī (<))
sg:Ac -am -om -em -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -osio, -ī -osio, -ī (<) -(e|o)s -(e|o)is (<), -(e|i)s (<) -ous (<), -(e|o)s -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -(a|ā)i, ?(-ās) -oi -es -ēs (<) -oues (<) -s
pl:Ac -ans -ons -ens -ns ?(-ns)
pl:n -ā -a -(a|ā)
pl:G -āsom -om -om ?(-som)
pl:D/Ab/L -ais -ois -(f|β)os ?(-(f|β)os)

Proto-Latino-Faliscan or Latin-only convergence of endings in -a/-ā

All previous endings in -a and/or -ā converge to -a. The paths to this differ between Sihler and Meiser. For Sihler, endings in -ā at first prevail over endings in -a (a-stem singular vocative turns to -ā under the influence of the nominative (p. 268), and the consonantic plural neuter under the influence of the o-stem (p. 263)), and later ("[r]elatively late in the history of [Latin]", but "predat[ing] the loss of final -d") all word-final -ā turn to -a, due to "iambic shortening", the influence of the shortened -am ending, or a mysteriously singular vowel shift (p. 78-79, 266-267). Meiser cannot accept such a vowel shift (p. 77) and proposes an inverse generalization of -a endings prevailing over -ā endings (p. 132, 136; see also Klein/Joseph/Fritz p. 756; Weiss p. 232)). Whether this convergence happened to the whole Latino-Faliscan branch or only Latin is unclear due to lack of vowel-length hints in the available Faliscan texts.

a o cons. -i- -u- -ē-
&c -i- -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -s- -n(d|t)i- -ti- &c
sg:N -a -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–) -∅ (–) -ss ?(–) (<<), -s -ss (<<), -s -s -s -s
sg:V -e (-e)|(-ī (<))
sg:Ac -am -om -em -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -osio, -ī -osio, -ī (<) -(e|o)s -(e|o)is (<), -(e|i)s (<) -ous (<), -(e|o)s -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -(a|ā)i, ?(-ās) -oi -es -ēs (<) -oues (<) -s
pl:Ac -ans -ons -ens -ns ?(-ns)
pl:n -a
pl:G -āsom -om -om ?(-som)
pl:D/Ab/L -ais -ois -(f|β)os ?(-(f|β)os)

Earliest Latin sound change of athematic plural dative/ablative

The -(f|β)- fricative in athematic dative/ablative endings changes to the voiced bilabial stop -b- (Sihler p. 147; Fortson p. 283; Meiser p. 101-103; Weiss p. 149).

a o cons. -i- -u- -ē-
&c -i- -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -s- -n(d|t)i- -ti- &c
sg:N -a -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–) -∅ (–) -ss ?(–) (<<), -s -ss (<<), -s -s -s -s
sg:V -e (-e)|(-ī (<))
sg:Ac -am -om -em -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -osio, -ī -osio, -ī (<) -(e|o)s -(e|o)is (<), -(e|i)s (<) -ous (<), -(e|o)s -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -(a|ā)i, ?(-ās) -oi -es -ēs (<) -oues (<) -s
pl:Ac -ans -ons -ens -ns ?(-ns)
pl:n -a
pl:G -āsom -om -om ?(-som)
pl:D/Ab/L -ais -ois -bos ?(-bos)

Earliest Latin partial fusion of athematic endings

The suffix-ending combination -i-bos, previously found in the plural dative/ablative of i-stems, encroaches upon the consonantal stems (and the u-stems, see below), probably as an import from there. Sihler p. 286 though argues that the new ending could just as well be an import from the u-stem: Both -i- and -u- here weakened early to an in-between sound that would only in writing be locked into either -i- or -u- again. In any case, this sound change is deemed the main reason -i-bos now also turns up for u-stems, with -u-bos attested rarely and not particularly early, therefore probably by later analogy or need for differentiation rather than as continuation of original -u-bos (Sihler p. 326; Weiss p. 252). More securly, the i-stem nominative/vocative plural ending -ēs encroaches upon the consontal stem, and the previous i-stem singular genitive ending -(e|o)is is completely pushed out by the -(e|i)s ending that is similar to the consonantal ending.

a o cons. -i- -u- -ē-
&c -i- -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -s- -n(d|t)i- -ti- &c
sg:N -a -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–) -∅ (–) -ss ?(–) (<<), -s -ss (<<), -s -s -s -s
sg:V -e (-e)|(-ī (<))
sg:Ac -am -om -em -em (<), -m -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -osio, -ī -osio, -ī (<) -(e|o)s -(e|i)s (<) -ous (<), -(e|o)s -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -(a|ā)i, ?(-ās) -oi -ēs -ēs (<) -oues (<) -s
pl:Ac -ans -ons -ens -ns ?(-ns)
pl:n -a
pl:G -āsom -om -om ?(-som)
pl:D/Ab/L -ais -ois -ibos -bos -ibos (<), ?(-bos) ?(-bos)

Earliest Latin plural accusatives

The various accusative plural -ns endings transform to -s with the preceding vowel lengthened (Sihler p. 215). For the a-stem, this may have happened much earlier – in Proto-Indo-European even, see Weiss p. 235-236, Meiser p. 133. For a similar interpretation regarding the o-stems, see Meiser p. 136, Weiss p. 206, and Klein/Joseph/Fritz p. 2084. On both suggestions, also consider Sihler p. 229, 254.

a o cons. -i- -u- -ē-
&c -i- -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -s- -n(d|t)i- -ti- &c
sg:N -a -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–) -∅ (–) -ss ?(–) (<<), -s -ss (<<), -s -s -s -s
sg:V -e (-e)|(-ī (<))
sg:Ac -am -om -em -em (<), -m -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -osio, -ī -osio, -ī (<) -(e|o)s -(e|i)s (<) -ous (<), -(e|o)s -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i -i -ei (<) -(ē|ō)u (<)
pl:N/V -(a|ā)i, ?(-ās) -oi -ēs -ēs (<) -oues (<) -s
pl:Ac -ās -ōs -ēs -īs (<) -ūs (<) -s
pl:n -a
pl:G -āsom -om -om ?(-som)
pl:D/Ab/L -ais -ois -ibos -bos -ibos (<), ?(-bos) ?(-bos)

Earliest (?) Latin loss of athematic singular locatives

The athematic singular locatives seem to mostly disappear in Latin, with only a few traces either in adverbs (only "diū" for the u-stem; a few more, like "rūrī" and "temperī", for the i-stem) or the consonantal ablative later on (Meiser p. 138-139, 147; Weiss p. 245, 252; Klein/Joseph/Fritz p. 761).

a o cons. -i- -u- -ē-
&c -i- -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -s- -n(d|t)i- -ti- &c
sg:N -a -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–) -∅ (–) -ss ?(–) (<<), -s -ss (<<), -s -s -s -s
sg:V -e (-e)|(-ī (<))
sg:Ac -am -om -em -em (<), -m -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -osio, -ī -osio, -ī (<) -(e|o)s -(e|i)s (<) -ous (<), -(e|o)s -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē), ?(-i) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i ?(-i) ?(-ei (<))
pl:N/V -(a|ā)i, ?(-ās) -oi -ēs -ēs (<) -oues (<) -s
pl:Ac -ās -ōs -ēs -īs (<) -ūs (<) -s
pl:n -a
pl:G -āsom -om -om ?(-som)
pl:D/Ab/L -ais -ois -ibos -bos -ibos (<), ?(-bos) ?(-bos)

Old Latin loss of -osio genitive

The o-stem singular genitive in -osio is barely found in pre-literary Latin (least controversial would be the sparse Lapis Satricanus inscription from the 6th or 5th century BCE), and by the 3rd century BCE seems to be replaced completely by -ī.

a o cons. -i- -u- -ē-
&c -i- -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -s- -n(d|t)i- -ti- &c
sg:N -a -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–) -∅ (–) -ss ?(–) (<<), -s -ss (<<), -s -s -s -s
sg:V -e (-e)|(-ī (<))
sg:Ac -am -om -em -em (<), -m -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -ī -ī (<) -(e|o)s -(e|i)s (<) -ous (<), -(e|o)s -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē), ?(-i) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i ?(-i) ?(-ei (<))
pl:N/V -(a|ā)i, ?(-ās) -oi -ēs -ēs (<) -oues (<) -s
pl:Ac -ās -ōs -ēs -īs (<) -ūs (<) -s
pl:n -a
pl:G -āsom -om -om ?(-som)
pl:D/Ab/L -ais -ois -ibos -bos -ibos (<), ?(-bos) ?(-bos)

Old Latin syncope of -os, -es, -e after -r-

The singular nominative and vocative endings -os and -e syncopate after -r- – certainly if -r- be post-consonantal, in which case -r- would become syllabic (-kros to -kr to -ker), but also in many cases if -r- be post-vocalic (see "puer", "vir", yet adjectives "ferus", "sincērus"). The timeline is vague, but a 500 BC inscription still has SAKROS, while post-vocalic -ros from earlier -sos via rhotacism never syncopates. Plautus has syncopated and un-syncopated forms, though the latter might be analogical restorations. A similar syncopic tendency is rarely expressed for -lo- stems (Ennius using "famul" instead of "famulus"; see Meiser p. 133, Weiss p. 221, Leumann p. 92). Similarly, post-consonantal -ri-s in i-stems syncopates to -r, i.e. -er (Leumann p. 98; Meiser p. 73; Weiss p. 243; Sihler p. 315). Furthermore, neuter -r(e|o)s-stem syncopates so that original *far(e|o)s- becomes *fars- (Sihler p. 306), assimilating (Meiser p. 116) to farr- (like genitive "farris", with the nominative simplifying the -rr-ending to "far"; Sihler p. 211).

a o cons. -i- -u- -ē-
-i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -s- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -er (<) -∅, -os -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–) -∅ (–) -ss ?(–) (<<), -s -ss (<<), -s -er (<<) -s -s -s
sg:V (-e)|(-ī (<)) -∅, -e -e
sg:Ac -am -om -em -em (<), -m -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-) -∅ -∅ (-∅)|(-ū (<))
sg:G -ās -ī (<) -ī -(e|o)s -(e|i)s (<) -ous (<), -(e|o)s -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē), ?(-i) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i ?(-i) ?(-ei (<))
pl:N/V -(a|ā)i, ?(-ās) -oi -ēs -ēs (<) -oues (<) -s
pl:Ac -ās -ōs -ēs -īs (<) -ūs (<) -s
pl:n -a
pl:G -āsom -om -om ?(-som)
pl:D/Ab/L -ais -ois -ibos -bos -ibos (<), ?(-bos) ?(-bos)

Old Latin rhotacism

By the 4th century BCE, any older intervocalic -s- has become -r- (Meiser p. 95; Sihler p. 171-173; Weiss p. 150-152; Klein/Joseph/Fritz p. 750).

a o cons. -i- -u- -ē-
-i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -er (<) -∅, -os -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–), -s (<, –) -ss ?(–) (<<), -s -ss (<<), -s -er (<<) -s -s -s
sg:V (-e)|(-ī (<)) -∅, -e -e
sg:Ac -am -om -em -em (<), -m -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -∅ (-∅)|(-ū (<))
sg:G -ās -ī (<) -ī -(e|o)s -(e|i)s (<) -ous (<), -(e|o)s -s
sg:Ab -ād -ōd ?(-ēd), ?(-ē), ?(-i) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i ?(-i) ?(-ei (<))
pl:N/V -(a|ā)i, ?(-ās) -oi -ēs -ēs (<) -oues (<) -s
pl:Ac -ās -ōs -ēs -īs (<) -ūs (<) -s
pl:n -a
pl:G -ārom -om -om ?(-rom)
pl:D/Ab/L -ais -ois -ibos -bos -ibos (<), ?(-bos) ?(-bos)

Old Latin a- and e-declension singular genitive in -ī

During the 3rd century BCE (but see Weiss p. 234 for a possible 4th century BCE example), except for some fossilized forms like "pater familiās", the a-declension singular genitive is re-shaped by analogy with the o-stem, with -ī appended to -ā-. The ē-declension's singular genitive, similar to and probably already in its earlier form derived from the a-declension, follows this change, though in rare cases -s seems to be continued (see Leumann p. 447 for disagreement).

a o cons. -i- -u- -ē-
-i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -er (<) -∅, -os -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–), -s (<, –) -ss ?(–) (<<), -s -ss (<<), -s -er (<<) -s -s -s
sg:V (-e)|(-ī (<)) -∅, -e -e
sg:Ac -am -om -em -em (<), -m -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -∅ (-∅)|(-ū (<))
sg:G -āī -ī (<) -ī -(e|o)s -(e|i)s (<) -ous (<), -(e|o)s -ī, ?(-s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē), ?(-i) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -(e|o)i ?(-i) ?(-ei (<))
pl:N/V -(a|ā)i, ?(-ās) -oi -ēs -ēs (<) -oues (<) -s
pl:Ac -ās -ōs -ēs -īs (<) -ūs (<) -s
pl:n -a
pl:G -ārom -om -om ?(-rom)
pl:D/Ab/L -ais -ois -ibos -bos -ibos (<), ?(-bos) ?(-bos)

Old Latin weakening of diphthongs

By the 3rd century BCE, former dipththongs -ai- and -oi- of short length have weakened to -ei- in various inscriptions. The previous state is attested only for the o-stem plural nominative/vocative and ablative/dative – in an old inscription from the 6th century BCE, and the probably similarly ancient Carmen Saliare. That the a-stem diphthongs in the singular locative and the plural nominative/vocative survive this stage (to turn to Classical -ae) seems to affirm their greater length – see Weiss p. 235 for a discussion. (Regular PIE outcome for the a-stem singular locative would have been a short -ai at least before some initial sounds of the following words, thus -āi here may have been a generalized sandhi variant; but Weiss also points to an alternate theory by Francisco Villar whereby short -ai may have survived weakening too).

a o cons. -i- -u- -ē-
-i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -er (<) -∅, -os -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–), -s (<, –) -ss ?(–) (<<), -s -ss (<<), -s -er (<<) -s -s -s
sg:V (-e)|(-ī (<)) -∅, -e -e
sg:Ac -am -om -em -em (<), -m -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -∅ (-∅)|(-ū (<))
sg:G -āī -ī (<) -ī -(e|o)s -(e|i)s (<) -ous (<), -(e|o)s -ī, ?(-s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē), ?(-i) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -ei ?(-i) ?(-ei (<))
pl:N/V -(a|ā)i, ?(-ās) -ei -ēs -ēs (<) -oues (<) -s
pl:Ac -ās -ōs -ēs -īs (<) -ūs (<) -s
pl:n -a
pl:G -ārom -om -om ?(-rom)
pl:D/Ab/L -eis -eis -ibos -bos -ibos (<), ?(-bos) ?(-bos)

Old Latin u-stem genitive singular

The u-stem singular genitive exhibits a variety of forms up until the 2nd century BCE. Some continue the u-stem-original -ous and the consonantal-stem-influenced -u(e|o)s (as -uos in the Senatus Consultum de Bacchanalibus 187 BCE, and as -uis in Terence). But in Plautus mostly, and some inscriptions, is found -ī, in analogy to the o-declension.

a o cons. -i- -u- -ē-
-i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -er (<) -∅, -os -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–), -s (<, –) -ss ?(–) (<<), -s -ss (<<), -s -er (<<) -s -s -s
sg:V (-e)|(-ī (<)) -∅, -e -e
sg:Ac -am -om -em -em (<), -m -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -∅ (-∅)|(-ū (<))
sg:G -āī -ī (<) -ī -(e|o)s -(e|i)s (<) -ī, -ous (<), -os, -is -ī, ?(-s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē), ?(-i) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ōi -ei -ei (<) -ouei (<), -ou (<) ?(-i), ?(-ei)
sg:L -(a|ā)i -ei ?(-i) ?(-ei (<))
pl:N/V -(a|ā)i, ?(-ās) -ei -ēs -ēs (<) -oues (<) -s
pl:Ac -ās -ōs -ēs -īs (<) -ūs (<) -s
pl:n -a
pl:G -ārom -om -om ?(-rom)
pl:D/Ab/L -eis -eis -ibos -bos -ibos (<), ?(-bos) ?(-bos)

Old Latin initial stage of shortening and monophthongization of diphthongs

By the 3rd century BCE, short diphthong -ei- probably monophthongizes to a somehwat higher -ē- (Meiser p. 58, 70-72 and Weiss p. 101, 143 argue for this; Sihler p. 53 allows for it, but calls the evidence "not rigorous"). Less epigraphic evidence is found for a monophthongization of -ou- to -ō- (Meiser p. 49; Weiss p. 103; not argued by Sihler). For quite a while, -EI- and -E- are found to interchangeably represent both the diphthong and the single vowel its monophthongization points to, i.e. EI is even found where -ē- (or later -ī-) would have no ancestry in -ei-; the situation seems to be similar with -OU- and -O-. Furthermore, by the 3rd century BCE at latest, the long diphthong ending -ōi of the o-stem singular dative (still attested for the 6th century BCE) has lost its final glide, possibly (Meiser p. 72-73) due to a generalized Sandhi mechanism. The ē-declension's singular genitive and dative are depicted in possible intermediary stages between the attested outcome in Plautus, Terence etc., and their probable initial morphological construction (for the dative, either appending regular -ei to the stem's -ē-, or -ē-i by analogy with the a-declension's -ā-i; for the genitive, -ē-ī in analoge to the a-declension's -ā-ī).

a o cons. -i- -u- -ē-
-i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -er (<) -∅, -os -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–), -s (<, –) -ss ?(–) (<<), -s -ss (<<), -s -er (<<) -s -s -s
sg:V (-e)|(-ī (<)) -∅, -e -e
sg:Ac -am -om -em -em (<), -m -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -∅ (-∅)|(-ū (<))
sg:G -āī -ī (<) -ī -(e|o)s -(e|i)s (<) -ī, -(ou|ō)s (<), -os, -is ?(-ē|(ei)) (<)), ?(-(e|ē)ī (<)), ?(-s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē), ?(-i) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ō -ē -ē (<) -(ou|ō|u)ē (<), -(ou|ō) (<) ?(-ē (<)) ?(-(e|ē)i (<))
sg:L -(a|ā)i -ē ?(-i) ?(-ē (<))
pl:N/V -(a|ā)i, ?(-ās) -ē -ēs -ēs (<) -(ou|ō)es (<) -s
pl:Ac -ās -ōs -ēs -īs (<) -ūs (<) -s
pl:n -a
pl:G -ārom -om -om ?(-rom)
pl:D/Ab/L -ēs -ēs -ibos -bos -ibos (<), ?(-bos) ?(-bos)

Old Latin o-stem plural genitive

From the 3rd century BCE on, an o-stem plural genitive in -ōrom (later -ōrum) is attested, which will more and more (but never completely) supplant original -om (later -um). It most probably derives from analogy with the a-declension.

a o cons. -i- -u- -ē-
-i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -er (<) -∅, -os -os -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–), -s (<, –) -ss ?(–) (<<), -s -ss (<<), -s -er (<<) -s -s -s
sg:V (-e)|(-ī (<)) -∅, -e -e
sg:Ac -am -om -em -em (<), -m -m -em (<)
sg:n -om -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -∅ (-∅)|(-ū (<))
sg:G -āī -ī (<) -ī -(e|o)s -(e|i)s (<) -ī, -(ou|ō)s (<), -os, -is ?(-ē|(ei)) (<)), ?(-(e|ē)ī (<)), ?(-s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē), ?(-i) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ō -ē -ē (<) -(ou|ō|u)ē (<), -(ou|ō) (<) ?(-ē (<)) ?(-(e|ē)i (<))
sg:L -(a|ā)i -ē ?(-i) ?(-ē (<))
pl:N/V -(a|ā)i, ?(-ās) -ē -ēs -ēs (<) -(ou|ō)es (<) -s
pl:Ac -ās -ōs -ēs -īs (<) -ūs (<) -s
pl:n -a
pl:G -ārom -ōrom, -om -om ?(-rom)
pl:D/Ab/L -ēs -ēs -ibos -bos -ibos (<), ?(-bos) ?(-bos)

Old Latin vowel reductions

By the 3rd century BCE, in closed final syllables, short -o- becomes -u-, except after -(u|v)- (Meiser p. 71, 84; Sihler p. 66), and around the same time, short -e- in final closed syllables becomes -i-, except before nasals (Meiser p. 71; Sihler p. 65; Weiss p. 138 sees an early example around 350 BCE already). The io-stem singular vocative definitely contracts to -ī, except "als grammatische Normierung oder gräzisierend in stilisierter alter Dichtung" (Leumann p. 424), as in Livius Andronicus.

a o cons. -i- -u- -ē-
-u- -i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -us -er (<) -∅, -us -us -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–), -s (<, –) -ss ?(–) (<<), -s -ss (<<), -s -er (<<) -s -s -s
sg:V -e -ī (<) -∅, -e -e
sg:Ac -am -om -um -em -em (<), -m -m -em (<)
sg:n -om -um -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -∅ (-∅)|(-ū (<))
sg:G -āī -ī -ī (<) -ī -(i|u)s -s -ī, -(ou|ō)s (<), -os, -is ?(-ē|(ei)) (<)), ?(-(e|ē)ī (<)), ?(-s)
sg:Ab -ād -ōd ?(-ēd), ?(-ē), ?(-i) -īd (<) -ūd (<) -d, ?(-ē)
sg:D -āi -ō -ē -ē (<) -(ou|ō|u)ē (<), -(ou|ō) (<) ?(-ē (<)) ?(-(e|ē)i (<))
sg:L -(a|ā)i -ē ?(-i) ?(-ē (<))
pl:N/V -(a|ā)i, ?(-ās) -ē -ēs -ēs (<) -(ou|ō)es (<) -s
pl:Ac -ās -ōs -ēs -īs (<) -ūs (<) -s
pl:n -a
pl:G -ārum -ōrom, -om -ōrum, -um -um -um -om -rum
pl:D/Ab/L -ēs -ēs -ibus -bus -ibus (<), ?(-bus) -bus

Old Latin singular ablative and the loss of -d

During the 3rd century BCE, word-final -d after long vowels is lost in speech, though still used in writing until the 2nd century BCE (Meiser p. 100; Weiss p. 155; Sihler p. 228). This accounts for most singular ablatives of Classical Latin, except for the consonant-stem ending in short -e. A consonant-stem ending -ED is attested epigraphically in the 3rd century BCE, but may be a false archaism (Weiss p. 238), and no continuation in -ē is found (unless one follows Sihler p. 285 arguing for the possibility that Classical -e might be a shortened -ē, but with that -ē deriving straight from the PIE instrumental); Classical Latin -e instead is thought to derive from the locative ending in -i, but how and when that might have become the ablative ending remains unclear (Sihler p. 285).

a o cons. -i- -u- -ē-
-u- -i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -us -er (<) -∅, -us -us -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–), -s (<, –) -ss ?(–) (<<), -s -ss (<<), -s -er (<<) -s -s -s
sg:V -e -ī (<) -∅, -e -e
sg:Ac -am -om -um -em -em (<), -m -m -em (<)
sg:n -om -um -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -∅ (-∅)|(-ū (<))
sg:G -āī -ī -ī (<) -ī -(i|u)s -s -ī, -(ou|ō)s (<), -os, -is ?(-ē|(ei)) (<)), ?(-(e|ē)ī (<)), ?(-s)
sg:Ab -ā -ō ?(-i), ?(-ē) -ī (<) -ū (<) -∅
sg:D -āi -ō -ē -ē (<) -(ou|ō|u)ē (<), -(ou|ō) (<) ?(-ē (<)) ?(-(e|ē)i (<))
sg:L -(a|ā)i -ē ?(-i) ?(-ē (<))
pl:N/V -(a|ā)i, ?(-ās) -ē -ēs -ēs (<) -(ou|ō)es (<) -s
pl:Ac -ās -ōs -ēs -īs (<) -ūs (<) -s
pl:n -a
pl:G -ārum -ōrom, -om -ōrum, -um -um -um -om -rum
pl:D/Ab/L -ēs -ēs -ibus -bus -ibus (<), ?(-bus) -bus

Old to Classical Latin fusion of i-stem plural nominative and accusative

Into the Classical period, the consonant-stem accusative plural encroaches upon, and more and replaces, the i-stem ending. The parallelism of -ēs and -īs extends to the nominative plural (with many examples for -īs already in Plautus), probably in analogy to the formal identity of both cases in other declensions.

a o cons. -i- -u- -ē-
-u- -i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -us -er (<) -∅, -us -us -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–), -s (<, –) -ss ?(–) (<<), -s -ss (<<), -s -er (<<) -s -s -s
sg:V -e -ī (<) -∅, -e -e
sg:Ac -am -om -um -em -em (<), -m -m -em (<)
sg:n -om -um -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -∅ (-∅)|(-ū (<))
sg:G -āī -ī -ī (<) -ī -(i|u)s -s -ī, -(ou|ō)s (<), -os, -is ?(-ē|(ei)) (<)), ?(-(e|ē)ī (<)), ?(-s)
sg:Ab -ā -ō ?(-i), ?(-ē) -ī (<) -ū (<) -∅
sg:D -āi -ō -ē -ē (<) -(ou|ō|u)ē (<), -(ou|ō) (<) ?(-ē (<)) ?(-(e|ē)i (<))
sg:L -(a|ā)i -ē ?(-i) ?(-ē (<))
pl:N/V -(a|ā)i, ?(-ās) -ē -ēs -īs (<), -ēs (<) -(ou|ō)es (<) -s
pl:Ac -ās -ōs -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:n -a
pl:G -ārum -ōrom, -om -ōrum, -um -um -um -om -rum
pl:D/Ab/L -ēs -ēs -ibus -bus -ibus (<), ?(-bus) -bus

Old Latin later stage of shortening and monophthongization of diphthongs

Between the end of the 3rd century and the middle of the 2nd century BCE, the monophthongization of former -ei- has (after the possible intermediate stage of -ē-) developed to -ī-, and that of former -ou- to -ū- (after the possible intermediate stage of -ō-). The u-stem plural nominative/vocative ending -ūs may emerge from this with questionable syncope of a previous e-ablaut ending (from -ou-es); an alternate theory is that the accusative ending was used here in analogy to the nominative-accusative similarity among consonantal stems (Meiser p. 146; Sihler p. 325-326; Weiss p. 124, 252). The u-stem singular dative is depicted here in its final/Classical form as having a short -u- before the glide, which fits regular shortening of vowels before other vowels in Latin (Sihler p. 80) in case -ouei might have turned to something like -ūī in between. The singular genitive of the a-declension shortens to -ai; again, its ē-declension sibling seems to follow this development, but to what distribution of -ei, -ēī etc. precisely is a bit unclear; the same confusion holds true for the singular dative; a short diphthong -ei was probably preserved in monosyllabic words like "rei"; the dative in -ē might through loss of final glide directly continue -ēi rather than -ei (Weiss p. 255).

a o cons. -i- -u- -ē-
-u- -i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -us -er (<) -∅, -us -us -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–), -s (<, –) -ss ?(–) (<<), -s -ss (<<), -s -er (<<) -s -s -s
sg:V -e -ī (<) -∅, -e -e
sg:Ac -am -om -um -em -em (<), -m -m -em (<)
sg:n -om -um -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -∅ (-∅)|(-ū (<))
sg:G -ai -ī -ī (<) -ī -(i|u)s -s -ī, -ūs (<), -os, -is -(ī|ē) (<), -ei (<), -(e|ē)ī (<), ?(-s)
sg:Ab -ā -ō ?(-i), ?(-ē) -ī (<) -ū (<) -∅
sg:D -āi -ō -ī -ī (<) -ī, -ū (<) -(ī|ē) (<), -ei (<)
sg:L -(a|ā)i -ī ?(-i) ?(-ī (<))
pl:N/V -(a|ā)i, ?(-ās) -ī -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:Ac -ās -ōs -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:n -a
pl:G -ārum -ōrom, -om -ōrum, -um -um -um -om -rum
pl:D/Ab/L -īs -īs -ibus -bus -ibus (<), ?(-bus) -bus

2nd century BCE endings in -e or -ae

By the 2nd century BCE, word final short -i has become -e (Sihler p. 65; Meiser p. 72; Weiss p. 146), and even in the a-stem diphthong endings, where the final glide is now written as such (Weiss p. 103 for how this might indicate not just an epigraphic but a phonetic change). There is no indication for the -AE that emerges to describe a long diphthong -āe, and regularly any vowel would have shortened before another in Latin (Sihler p. 80, so Classical -ae is assumed from here on. The consonant-stem ablative in -e most plausibly continues an -i derived from the locative (but see Sihler p. 285 for alternate interpretations), and encroaches upon the i-stem declension.

a o cons. -i- -u- -ē-
-u- -i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -us -er (<) -∅, -us -us -s -(b|p)s (<) -x (<) -ss (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–), -s (<, –) -ss ?(–) (<<), -s -ss (<<), -s -er (<<) -s -s -s
sg:V -e -ī (<) -∅, -e -e
sg:Ac -am -om -um -em -em (<), -m -m -em (<)
sg:n -om -um -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -e (<) (-∅)|(-ū (<))
sg:G -ae -ī -ī (<) -ī -(i|u)s -s -ī, -ūs (<), -os, -is -(ī|ē) (<), -ei (<), -(e|ē)ī (<), ?(-s)
sg:Ab -ā -ō -e -e (<), -ī (<) -ū (<) -∅
sg:D -ae -ō -ī -ī (<) -ī, -ū (<) -(ī|ē) (<), -ei (<)
sg:L -ae -ī ?(-e) ?(-ī (<))
pl:N/V -ae, ?(-ās) -ī -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:Ac -ās -ōs -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:n -a
pl:G -ārum -ōrom, -om -ōrum, -um -um -um -om -rum
pl:D/Ab/L -īs -īs -ibus -bus -ibus (<), ?(-bus) -bus

2nd century BCE reduction of -ss to -s

Word-final -s-s (from -t-s) simplifies to -s (Plautus still has "miless" with final syllable scanning, but Ennius already has "miles" with final syllable scanning short – Leumann p. 222, Meiser p. 113-114).

a o cons. -i- -u- -ē-
-u- -i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -us -er (<) -∅, -us -us -s -(b|p)s (<) -x (<) -s (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–), -s (<, –) -s (–) (<<), -s -s (<<), -s -er (<<) -s -s -s
sg:V -e -ī (<) -∅, -e -e
sg:Ac -am -om -um -em -em (<), -m -m -em (<)
sg:n -om -um -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -e (<) (-∅)|(-ū (<))
sg:G -ae -ī -ī (<) -ī -(i|u)s -s -ī, -ūs (<), -os, -is -(ī|ē) (<), -ei (<), -(e|ē)ī (<), ?(-s)
sg:Ab -ā -ō -e -e (<), -ī (<) -ū (<) -∅
sg:D -ae -ō -ī -ī (<) -ī, -ū (<) -(ī|ē) (<), -ei (<)
sg:L -ae -ī ?(-e) ?(-ī (<))
pl:N/V -ae, ?(-ās) -ī -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:Ac -ās -ōs -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:n -a
pl:G -ārum -ōrom, -om -ōrum, -um -um -um -om -rum
pl:D/Ab/L -īs -īs -ibus -bus -ibus (<), ?(-bus) -bus

2nd century BCE loss of singular -e in some i-stem neuters

By the 2nd century BCE (after -i had become -e, but with an example already in the Senatus consultum de Bacchanalibus – see Leumann p. 92), i-stem neuters in the singular nominative/accusative lose their ending after polysyllabic stems in -ā(r|l)-. Notably, this does not extend to adjectives (Weiss p. 147; p. 352). While Sihler p. 69, 316 sees this change only affect polysyllabic stems, Meiser p. 74 calls the -e in mar-e "restituiert".

a o cons. -i- -u- -ē-
-u- -i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -us -er (<) -∅, -us -us -s -(b|p)s (<) -x (<) -s (<) -∅ (–) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (–), -s (<, –) -s (–) (<<), -s -s (<<), -s -er (<<) -s -s -s
sg:V -e -ī (<) -∅, -e -e
sg:Ac -am -om -um -em -em (<), -m -m -em (<)
sg:n -om -um -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -e (<), -∅ (<) (-∅)|(-ū (<))
sg:G -ae -ī -ī (<) -ī -(i|u)s -s -ī, -ūs (<), -os, -is -(ī|ē) (<), -ei (<), -(e|ē)ī (<), ?(-s)
sg:Ab -ā -ō -e -e (<), -ī (<) -ū (<) -∅
sg:D -ae -ō -ī -ī (<) -ī, -ū (<) -(ī|ē) (<), -ei (<)
sg:L -ae -ī ?(-e) ?(-ī (<))
pl:N/V -ae, ?(-ās) -ī -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:Ac -ās -ōs -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:n -a
pl:G -ārum -ōrom, -om -ōrum, -um -um -um -om -rum
pl:D/Ab/L -īs -īs -ibus -bus -ibus (<), ?(-bus) -bus

2nd century BCE shortening of vowels before -l and -r

By the 2nd century BCE, in polysyllables, vowels in final syllables before -l and -r are shortened (Sihler p. 79; Weiss p. 128; Meiser p. 77; Leumann p. 111).

a o cons. -i- -u- -ē-
-u- -i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -us -er (<) -∅, -us -us -s -(b|p)s (<) -x (<) -s (<) -∅ (v for poly-syl.) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (– for mono-, v for poly-syl.), -s (<, –) -s (–) (<<), -s -s (<<), -s -er (<<) -s -s -s
sg:V -e -ī (<) -∅, -e -e
sg:Ac -am -om -um -em -em (<), -m -m -em (<)
sg:n -om -um -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -e (<), -∅ (<, v) (-∅)|(-ū (<))
sg:G -ae -ī -ī (<) -ī -(i|u)s -s -ī, -ūs (<), -os, -is -(ī|ē) (<), -ei (<), -(e|ē)ī (<), ?(-s)
sg:Ab -ā -ō -e -e (<), -ī (<) -ū (<) -∅
sg:D -ae -ō -ī -ī (<) -ī, -ū (<) -(ī|ē) (<), -ei (<)
sg:L -ae -ī ?(-e) ?(-ī (<))
pl:N/V -ae, ?(-ās) -ī -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:Ac -ās -ōs -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:n -a
pl:G -ārum -ōrom, -om -ōrum, -um -um -um -om -rum
pl:D/Ab/L -īs -īs -ibus -bus -ibus (<), ?(-bus) -bus

2nd to 1st century BCE -ābus and -ubus endings

From the 2nd century BCE on, analogical a-stem plural dative/ablative forms in -ābus are found next to regular -īs, plausibly to differentiate a-stem forms from o-stem ones. Similarly, a u-stem plural dative/ablative occasionally appears besides regular -ibus, which may continue an ancient regular form in -u-bus, or be a more recent analogy to differentiate the otherwise identical plural dative-ablatives of words like "arx" and "arcus"; in any case, it is safely attested from the 1st century BCE on.

a o cons. -i- -u- -ē-
-u- -i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -os -us -er (<) -∅, -us -us -s -(b|p)s (<) -x (<) -s (<) -∅ (v for poly-syl.) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (– for mono-, v for poly-syl.), -s (<, –) -s (–) (<<), -s -s (<<), -s -er (<<) -s -s -s
sg:V -e -ī (<) -∅, -e -e
sg:Ac -am -om -um -em -em (<), -m -m -em (<)
sg:n -om -um -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -e (<), -∅ (<, v) (-∅)|(-ū (<))
sg:G -ae -ī -ī (<) -ī -(i|u)s -s -ī, -ūs (<), -os, -is -(ī|ē) (<), -ei (<), -(e|ē)ī (<), ?(-s)
sg:Ab -ā -ō -e -e (<), -ī (<) -ū (<) -∅
sg:D -ae -ō -ī -ī (<) -ī, -ū (<) -(ī|ē) (<), -ei (<)
sg:L -ae -ī ?(-e) ?(-ī (<))
pl:N/V -ae, ?(-ās) -ī -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:Ac -ās -ōs -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:n -a
pl:G -ārum -ōrom, -om -ōrum, -um -um -um -om -rum
pl:D/Ab/L -īs, -ābus -īs -ibus -bus -bus, -ibus (<) -bus

1st century BCE vowel reductions

Around the end of the Republic, from the second half of the 1st century BCE on, earlier -o- becomes -u- even after -(u|v)- (Sihler p. 66, Meiser p. 84).

a o cons. -i- -u- -ē-
-i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -us -er (<) -∅, -us -us -s -(b|p)s (<) -x (<) -s (<) -∅ (v for poly-syl.) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (– for mono-, v for poly-syl.), -s (<, –) -s (–) (<<), -s -s (<<), -s -er (<<) -s -s -s
sg:V (-e)|(-ī (<)) -∅, -e -e
sg:Ac -am -um -em -em (<), -m -m -em (<)
sg:n -um -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -e (<), -∅ (<, v) (-∅)|(-ū (<))
sg:G -ae -ī (<) -ī -(i|u)s -s -ī, -ūs (<), -os, -is -(ī|ē) (<), -ei (<), -(e|ē)ī (<), ?(-s)
sg:Ab -ā -ō -e -e (<), -ī (<) -ū (<) -∅
sg:D -ae -ō -ī -ī (<) -ī, -ū (<) -(ī|ē) (<), -ei (<)
sg:L -ae -ī ?(-e) ?(-ī (<))
pl:N/V -ae, ?(-ās) -ī -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:Ac -ās -ōs -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:n -a
pl:G -ārum -ōrum, -um -um -um -um -rum
pl:D/Ab/L -īs, -ābus -īs -ibus -bus -bus, -ibus (<) -bus

1st century BCE fusion of ē-stem singular genitive and dative

Around the 1st century BCE, the similarity of the ē-stem's singular dative and genitive endings levels their remaining difference, i.e. the dative now may also scan as a long diphthong ending in -ī in authors such as Lucretius and Horace. A rule whereby the diphthong would be shorter -eī after consonants (fideī) but longer -ēī after -i- (diēī) is declared by Classical grammarians later on, but may just be "an artifact of schoolmasterly prescriptivism" (Sihler p. 342); except for monosyllables like "rei", the actual pronounciation of the ending was at this time probably mere -ī (from -ei) (Leumann p. 446-447).

a o cons. -i- -u- -ē-
-i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -us -er (<) -∅, -us -us -s -(b|p)s (<) -x (<) -s (<) -∅ (v for poly-syl.) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (– for mono-, v for poly-syl.), -s (<, –) -s (–) (<<), -s -s (<<), -s -er (<<) -s -s -s
sg:V (-e)|(-ī (<)) -∅, -e -e
sg:Ac -am -um -em -em (<), -m -m -em (<)
sg:n -um -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -e (<), -∅ (<, v) (-∅)|(-ū (<))
sg:G -ae -ī (<) -ī -(i|u)s -s -ī, -ūs (<), -os, -is -(ī|ē) (<), -ei (<), -(e|ē)ī (<)
sg:D -ae -ō -ī -ī (<) -ī, -ū (<)
sg:Ab -ā -ō -e -e (<), -ī (<) -ū (<) -∅
sg:L -ae -ī ?(-e) ?(-ī (<))
pl:N/V -ae, ?(-ās) -ī -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:Ac -ās -ōs -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:n -a
pl:G -ārum -ōrum, -um -um -um -um -rum
pl:D/Ab/L -īs, -ābus -īs -ibus -bus -bus, -ibus (<) -bus

1st century BCE o-stem ī-endings with/after -i-

In the 1st century BCE, for o-stems in -i-, next to the contracted -ī in the singular genitive, a non-contracted -i-ī appears; at the same time, the locative in -i-ī contracts to -ī.

a o cons. -i- -u- -ē-
-i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -us -er (<) -∅, -us -us -s -(b|p)s (<) -x (<) -s (<) -∅ (v for poly-syl.) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (– for mono-, v for poly-syl.), -s (<, –) -s (–) (<<), -s -s (<<), -s -er (<<) -s -s -s
sg:V (-e)|(-ī (<)) -∅, -e -e
sg:Ac -am -um -em -em (<), -m -m -em (<)
sg:n -um -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -e (<), -∅ (<, v) (-∅)|(-ū (<))
sg:G -ae -ī, -ī (<) -ī -(i|u)s -s -ī, -ūs (<), -os, -is -(ī|ē) (<), -ei (<), -(e|ē)ī (<)
sg:D -ae -ō -ī -ī (<) -ī, -ū (<)
sg:Ab -ā -ō -e -e (<), -ī (<) -ū (<) -∅
sg:L -ae -ī (<) -ī ?(-e) ?(-ī (<))
pl:N/V -ae, ?(-ās) -ī -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:Ac -ās -ōs -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:n -a
pl:G -ārum -ōrum, -um -um -um -um -rum
pl:D/Ab/L -īs, -ābus -īs -ibus -bus -bus, -ibus (<) -bus

Classical Latin loss and fossilization of older endings

Various endings seem to have disappeared in Classical Latin, or only survive in fossilized forms, such as the o-stem plural genitive in -um, or some u-stem singular genitive endings. [TODO: consonant locative singular in -ī is modeled in analogy to o-stems (Sihler" p. 285). NOT DEPICTED: u-stem (or rather: ū-stem) freak cases "sūs" and "grūs" (Sihler p. 328); freak case "Iuppiter"/"Iovis" (Sihler p. 338-339); freak case "bōs"/"bovis" (Sihler p. 334-336); i- and mixed-stem freak cases of "senex"/"senis", "urbs"/"urbis", "nox"/"noctis", and s-stem root nouns with genitive plural in -ium, and the opposite with nominative singular in -is and otherwise consonant-stem endings (Sihler p. 319); kinship nouns of animate consonant stems in -r with nominative singular in -er (<) (like "pater"/"patris") (Sihler p. 293); consonant stems in -n- freak cases of "iuvenis"/"iuvenis", "sangu(i|ī)s"/"sanguinis", "liēn"/"li(e|ē)nis" (Sihler p. 291, 295-296, 298); consonant stems in -l- freak case of "sāl"/"salis" (Sihler p. 304); root noun freak cases of "vīs" and "spēs"/"speī" (Sihler p. 311); neuter r/n-stems freak cases of "femur"/"feminis", "iecur"/"iocineris", "iter"/"itineris" Sihler p. 298-299); nouns in -ēs nominative singular, mostly behaving like i-stem nouns (Sihler p. 316)]

a o cons. -i- -u- -ē-
-i- -Cr- -Vr- &c -p- -b- -(g|c|v|h)- -C(d|t)- -V(d|t)- -l- -m- -n- -r- -n(d|t)i- -ti- -Cri- &c
sg:N -a -us -er (<) -∅, -us -us -s -(b|p)s (<) -x (<) -s (<) -∅ (v for poly-syl.) -?(p)s -ō (<<, < for stem in -rn-), -en (<<) -∅ (– for mono-, v for poly-syl.), -s (<, –) -s (–) (<<), -s -s (<<), -s -er (<<) -s -s -s
sg:V (-e)|(-ī (<)) -∅ -e
sg:Ac -am -um -em -em (<), -m -m -em (<)
sg:n -um -∅ (<) -∅ -∅ (< for stems in -ll-) -en (<<) -∅ (< for stems in -rr-), -s (<) -e (<), -∅ (<, v) (-∅)|(-ū (<))
sg:G -ae -ī, -ī (<) -ī -(i|u)s -s -ūs (<), -is -(ī|ē) (<), -ei (<), -(e|ē)ī (<)
sg:D -ae -ō -ī -ī (<) -ī, -ū (<)
sg:Ab -ā -ō -e -e (<), -ī (<) -ū (<) -∅
sg:L -ae -ī (<) -ī ?(-e) ?(-ī (<))
pl:N/V -ae, ?(-ās) -ī -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:Ac -ās -ōs -ēs -īs (<), -ēs (<) -ūs (<) -s
pl:n -a
pl:G -ārum -ōrum -um -um -um -rum
pl:D/Ab/L -īs, -ābus -īs -ibus -bus -bus, -ibus (<) -bus

Bibliography