1. Facts
  2. Sources of the data
  3. Orthography
  4. Verbal inflection
  5. Inflectional classes by person affix
  6. Inflectional classes by affix-mood prefix
  7. Inflectional classes by singular/plural stem split


Amuzgo is a small language family. There are two main languages: Xochistlahuacan Amuzgo and San Pedro Amuzgos Amuzgo. The database is from the language spoken in the village of San Pedro Amuzgos in the state of Oaxaca by about 3,500 speakers.

Sources of the data

The database includes data taken from two different lexical sources.

  • The first source of data was made available to us courtesy of Fermín Tapia, an elderly linguist from Mexico who is a native speaker of Amuzgo and who lives in the village of San Pedro Amuzgos. Fermín Tapia intends to publish these materials in print. In the meantime, he has given us his permission to publish them online as part of this database, and we are immensely grateful for this. The material includes the fully inflected forms of 499 verbs.

    Fermín Tapia's materials were passed on to us by Yuni Kim from the University of Manchester, and we are very thankful for this. Yuni Kim has done a wonderful job at combing Fermín Tapia's materials for typos. Verbs that remain untouched are signalled with a double asterisk (**).

  • The second source is from the SIL published dictionary:

    Stewart, Cloyd and Ruth D. Stewart (eds.). 2000. Diccionario Amuzgo de San Pedro Amuzgos, Oaxaca [Serie de vocabularios y diccionarios indígenas Mariano Silva y Aceves, 44]. Mexico City: Instituto Lingüístico de Verano.

We also made use of the following grammatical description:

Buck, Marjorie. 2000. Gramática amuzga de San Pedro Amuzgos, Oaxaca, in Cloyd Stewart and Ruth D. Stewart (eds.), Diccionario Amuzgo de San Pedro Amuzgos, Oaxaca [Serie de vocabularios y diccionarios indígenas Mariano Silva y Aceves, 44]. Mexico City: Instituto Lingüístico de Verano. pp. 361-480. Available online.


We have respected the orthography used in the sources, except where indicated below:

IPA notation
Source notation
Our notation
ɛ ë ɛ
ɔ ö ɔ
b v b
ʰC jC h
Cj Ch
h j h
ʃ x
ch ch
ʔ ʔ
k c/__a, u, o k
qu/__i, e
Cj Cy Cy
j y y
ɲ ñ ñ
Cliticization h


There are eight contrasting tones.

Data from: Kim, Yuni. Forthcoming. Tonal overwriting and inflectional exponence in Amuzgo, in Enrique L. Palancar and Jean-Léo Léonard (eds.), Tone and inflection: New facts under new perspectives. DeGruyter Mouton.

Tone Fermín Tapia's Buck's 2000 Example
Low 1 L n̥ɛ ‘zacate’ (dry maize cane)
Mid 3 M n̥ɛ ‘cheap’
High 5 H n̥ɛ ‘dust’
Low Ascending 12 LL n̥ɛ ‘wilderness’
Mid Ascending 34 MM n̥ɛ ‘jump (CPL.3SG)’
Mid-High Ascending 35 MH n̥ɛ ‘ripe’
Mid Descending 31 HL n̥ɛ ‘already’
High Descending 53 HM n̥ɛ ‘sell (CPL.3SG)’

NOTE: Besides tones marked with numbers, Fermín Tapia's notation of tone also includes diacritics on vowels, which we have left untouched in the database: á ‘high’, à ‘low’, â ‘descending’.

Verbal inflection

Amuzgo verbal inflection is extremely complex, and thus all the more interesting, but we can only attempt a rudimentary sketch here to cover the basics. Verbs inflect for various aspect-moods by means of prefixes: Completive (also called ‘Preterite’ or ‘Past’); Incompletive (also called ‘Present’); Subjunctive (a tense used in some dependent clauses); Irrealis (also called ‘Future’); Past Progressive and Future Progressive.

Forms for 3SG of n⁵na³ba³ndyè¹ ‘skip’
CPL hna³ba³ndyè¹ ‘I skipped’
INCPL ʔ⁵na³ba³ndyè¹ ‘I'm skipping’
SUBJV ki³na³ba³ndyè¹ ‘I may skip’
IRR n⁵na³ba³ndyè¹ ‘I'll skip’
PST.PRG to³na³ba³ndyè¹ ‘I was skipping’, ‘I used to skip’
FUT.PRG n⁵ngo³na³ba³ndyè¹ ‘I'll be skipping’

Inflectional classes by person affix

Verbs fall into at least two main classes (i.e., there may be more) attending to how they encode person. Many intransitive verbs, called ‘process’ and ‘impersonal’ verbs by Buck (2000) inflect for subject by means of the same markers of person that encode object. Others, called ‘active’ by Buck (2000) are verbs that inflect for person of subject by more complex means, which may involve tonal changes, glottalization, pronominalization, etc. We refer to the latter as Class A verbs, and to the former as Class B.

Class A Class B
'get wet'
INCPL 1SG ko³-ndáʔ⁵=ha³ ma³-na³ba⁵³+ndy‹ɛ›¹²
2SG ko³-ndáʔ⁵=ʔu³ ma³-na³ba-ʔ¹²+ndyè-ʔ¹
3SG ko³-ndáʔ⁵=én⁵ ʔ⁵-na³ba³+ndyè¹
1PL.EXCL ko³-ndáʔ⁵há⁵=á⁵ ko³-nàʔ¹ba¹²+ndy‹ɛ›¹²
1PL.INCL ko³-ndáʔ⁵=hâ³¹ ko³-nàʔ¹bâ³¹+ndyê³¹
2PL ko³-ndáʔ⁵=ʔo³ ko³-nàʔ¹bà¹+ndyè¹=oʔ³
3PL ko³-ndáʔ⁵=en³ ko³-nàʔ¹bà¹+ndyè¹=hó⁵


  • A Class A verb like ko³ndáʔ⁵én⁵ ‘get wet’ has the prefix ko³- for the INCPL in all persons.
  • Class B verbs have ko³- for the INCPL only in the plural, but ma³- for the 1SG and 2SG, and ʔ⁵- for the 3SG (which is rendered in Buck (2000) as i-).
  • Some verbs in the database take a Ø- prefix for all persons in the singular for the INCPL. Most of these verbs have stative semantics.
  • Class B verbs ending in oral or nasal /o/ or /e/ undergo vowel lowering in the forms for the 1st person singular and 1st person plural exclusive. For the SIL entries, this has been indicated under additional information.
  • Class B verbs receive a glottal for the 2SG person, and also in other persons by means of glottal metathesis. To handle this variance, Buck (2000) proposes different classes annotated with capital letters. The database provides this information as additional information for each entry. For more information about what these classes mean see Buck (2000).

Inflectional classes by aspect-mood prefix

We have already pointed out differences in the encoding of the INCPL across verbs. Similarly, verbs fall into different inflectional classes attending to the prefix set they select to realize the remaining aspect-moods. There is a high degree of morphophonological complexity. We do not provide a segmentation of Fermín Tapia's data.

The SIL source proposes organising verbs into various different classes called ‘conjugational models’, which are given here for convenience. The database includes information about class membership for each verb:

1 Ø- n- ki-
1d Ø- n- (t>d; tj>dj) ki-
2 h- n- ki-
2n h- (ʔndy>ntyʔ; ʔñ >ñʔ) n- ki-
3 s- (ts>Ø) n- ki-
4 ʃ- (ch>Ø) n- ki-
5 Ø ngi- ki-
6 t- nk- k-
6ty ty- nk- k-
7 t- nt- Ø-
8 t- (k/kw>Ø) n- Ø-
8ty ty- (k/ky/kw>Ø) n- Ø-
8g t- (k>Ø) n- (k>g) Ø-
8tyg ty- (k>Ø) n- (k>g) Ø-
9 tyi- nin- Ø-
10 to- non- Ø-
11 t- (m>u) n- ki-
12 t- (m>u; in>en) ng- (m>u; in>en) k- (m>u; in>en)
13 t- (m>Ø; in>e) ngw- (m>Ø; in>e) kw- (m>Ø; in>e)
14 t- (m>Ø; a>o) ng- (m>Ø; a>o) k- (m>Ø; a>o)
15 t- (m>Ø; an>on) nk- (m>Ø; an>on) k- (m>Ø; an>on)
16 t- (m>Ø) nkw- (m>Ø) kw- (m>Ø)
17 t- (b>u) ng- (b>u) k- (b>u)
17ty tj- (b>u) ng- (b>u) k- (b>u)
18 t- (b>Ø) nk- (b>Ø) k- (b>u)
19 t- (b>Ø) ng- (b>Ø) k- (b>u)
20 t- (b>Ø) nkw- (b>Ø) kw- (b>u)
21 t- (b>Ø) ngw- (b>Ø) kw- (b>u)
22 t- (b>Ø) nd- (b>Ø) kw- (b>u)
23 t- (nd>Ø) n- kw- (nd>Ø)
24 t- (nd>Ø) Ø- ki-

We have reduced this variance to six macro-classes. Class I is a default. Some other classes apply attending to the phonological onset of the stem of the verb. The database also contains information about the membership of verbs to these alternative classes:

I Model T- N- KI- Onset of stem
a. 1 T-CV > CV N-CV > nCV KI-CV > kiCV C=t, tɁ, ty, th
nTC, hnTC, hnd, hndy
nch, nchh, nchɁ
hn, hñ, nh, ñh
b. 2 T-NV > hNV N-NV > nNV KI-NV > kiNV N=n, nɁ, ñ, Ɂñ, ñɁ, nd, ndy, Ɂndy
c. 3 T-tsV > sV N-tsV > ntsV KI-tsV > kitsV ts, tsɁ
d. 4 T-tʃV > V N-tʃV > ntʃV KI-tʃV > kitʃV ch, chɁ
e. 11 T-mV > tuV N-mV > nmV KI-mV > kimV m, Ɂm
f. 8 T-KV > tV N-KV > nKV KI-KV > kV K=k, kɁ, kh, kʷ

II Model T- NGI- KI- Onset of stem
a. 5 T-TV > TV NGI-TV > ngiTV KI-TV > kiTV ch, chɁ, ts, tsɁ, t, tɁ, ty, th
b. 6 T-CV > tHV NGI-HV > nkHV KI-HV > kHV (H=h, Ɂ) hV, ɁV, yɁ
c. 7 T-KV > tKV NGI-KV > ntKV KI-KV > kV k, kɁ, kh, kʷ
d. 17/12 T-ɁbV > tuV NGI-ɁbV > nguV KI-ɁbV > kuV Ɂba, 'bi, ɁbeɁ, m
e. 19 T-bVɁCV > tVCV NGI-bVɁCV > ngVCV KI-bVɁCV > kVCV baɁCV

III Model TI- NIN- K- Onset of stem
a. 9 TI-nKV > tinKV NIN-nKV > ninnKV K-nKV > nKV nkV, nkyV, nkyɁV, nkhV
b. 10 TI-nKo > tonKo NIN-nKo > nonnKo K-nKo > nKo nko, nkyo, nkyɁo, nkho

IV Model TO- NGO- KO- Onset of stem
14/15 TO-CV > toCV NGO-CV > ngoCV KO-CV > koCV m, Ɂm, mɁ

V Model T- NGU- KU- Onset of stem
a. 21 T-bVɁCV > tVCV NGU-bVɁCV > ngwVɁCV KU-bVɁCV > kwVɁCV baɁCV, beɁCV
b. 20 T-bHV > tHV NGU-bHV > nkwHV KU-bHV > kwHV (H=h, Ɂ) bɁV, bhV, (Ɂbe)

VI Model T- NDI- KU- Onset of stem
a. 22 T-biɁCV > tiCV NDI-biɁCV > ndiɁCV KU-biɁCV > kwiɁCV biɁCV
b. 23 T-ndiɁCV > tiCV NDI-ndiɁCV > nndiɁCV KU-biɁCV > kwiɁCV ndiɁCV

Inflectional classes by singular/plural stem split

Many verbs in Amuzgo have two distinct stems: one is used for all persons in the singular; the other for all persons in the plural. This distinction exists regardless of aspect-mood. The stem variants may contrast only in segments, only in tone, or in both segments and tone, as illustrated below.

‘forget’ –tʔio¹tzʔon³ –tʔio¹nʔo³
‘whistle’ –tiu¹ –tiu³
‘urinate’ –chiuʔ³⁵ –ndyʔiu¹²

This is an important phenomenon for the inflection of verbs in Amuzgo and the database provides information about different aspects of how the distinction is realized:

  • The database includes information about the internal structure of both singular and plural stems and about the tone of base forms.
  • The database gives information regarding the different type of contrasts that we have detected involved in the formation of the singular and the plural stem.
  • The contrast between a singular stem and a plural stem often involves a change in the onset of the stem (commonly by having a different stem formative). This has consequences for the choice of aspect-mood prefix, as it can happen that verbs select a given prefix set for the singular, but another for the plural.
  • Consequently, the database gives information about class membership to the aspect-mood inflectional classes for the singular and the plural forms of each verb, independently. Note, however, that this is only given for the SIL data.