Morphology of Tibetan verb stems

Back to the main Tibetan language page.

---

Terminology

Verbs have four stems—conventionally called the Present (P), Past (A), Future (F), and Imperative (I) and cited in this order—though not all need be distinct. Case of verbs with 1, 2, 3, and 4 distinct stems exist.

Intransitive verbs are usually said to have only the first two stems, P and A (distinct or identical). Transitive verbs have all four, distinct or identical.

For the purpose of describing the verbal morphology that distinguishes these stems, we break the verb syllable into four parts:

• Prefix (P), which is the orthographic prefix;
• Main (M), which is the orthographic superscript, root letter, and subscript, of which only the root letter itself undergoes change;
• Vowel (V), which is the orthographic vowel diacritic or inherent a;
• Coda (C), which is the orthographic consonantal suffix with secondary suffix.

Example: In བསྒྲགས་ bsgrags,

• P is བ, b,
• M is སྒྲ, sGr,
• V is ཨ, inherent a, and
• C is གས, gs.

Descriptive approaches

Beyer (1992) gives a description of how the four stems of a given verb can be derived from a hypothetical underlying form, the root, by the application of complex but more or less regular rules. Hahn (2009), on the other hand gives a survey of the forms that are actually attested and the relationships that may hold between them. Here I attempt to summarise both and to show their essential agreement. (This work is not complete yet).

The derivational approach

We need to know the root form of the verb (which is hypothesized to represent an actual historical form), as well as the derivational paradigm it belongs to. We proceed in the following way:

• Apply the paradigmatic affixes and changes;
• Apply any applicable affix adjustment rules;
• Apply any applicable root adjustment rules.

These root forms will be displayed in BOLD CAPS.

Paradigms

For intransitive verbs, only one paradigm exists (IN), while any one of four models (T1–T4) may apply to an intransitive verb ("—" here represents the verb root):

paradigm	P	A	F	I
IN	འ—	—ས
T1	འ—(ད)	བ—ས	བ—	—ས (a>o) (asp)
T2	འ—(ད)	བ—ས	ག—	—ས (a>o) (asp)
T3	ག— (a>o)	བ—ས	བ—	—ས (a>o) (asp)
T4	ག— (a>o)	བ—ས	ག—	—ས (a>o) (asp)

XXXTESTstart

paradigm	P	A	F	I
IN	འ—	—ས
T1	འ—(ད)	བ—ས	བ—	—ས (a>o) (asp)
T2			ག—
T3	ག— (a>o)		བ—
T4			ག—

XXXTESTend

Here "(a>o)" represents rounding: a vowel a in the root becomes o.

"(asp)" represents aspiration: it changes nonaspirated oral stops (columns I and III) without prefixes into the homorganic column II oral stop. It has no effect on nasals, fricatives, aspirates, or consonants with a prefix

Affix adjustment rules

Since for a given verb the result of adding the affixes may be phonotactically illegal, under certain condition the individual prefixes are blocked (i.e. are replaced by zero) or dissimilated to a phonotactically acceptable consonant:

The prefix འ

The prefix འ- is blocked

• before a superscript or prefix to a root consonant: འ+RKU > རྐུ་
• before non-stop root consonants, e.g. འ+ZA > ཟ་
  • NB: In several verbs, འ+L > ལྡ, e.g. འ+LANG > ལྡང་
• before root nasal consonants, e.g. འ+NYO > ཉོ་

The prefix ག-

The prefix ག- is affected in the following ways:

• Before a superscript or prefix to a root consonant ག is blocked: ག+RTAG > རྟོག་
• ག becomes ད when the root consonant it is affixed is from rows 1 or 4 (that is, is velar or labial): ག+KROL > དཀྲོལ་
• and it appears as ག in front of root consonants from rows 2, 3, and 5 (palatal, dental, and dental affricates): ག+DUL > གདུལ་.
  • ?exceptions for affricates?

The prefix བ-

• is blocked before any labial root consonant: བ+BOR+ས > བོར་
• is blocked before nasals: བ+NYO+ས > ཉོས་
• is blocked before aspirated stops: བ+KHUR+ས > ཁུར་
  • ?exception for aspirated affricate or affricates generally?
• is blocked as a pre-prefix before a stop prefix: བ+_GNANGS_+ས > གནངས་ (nor does the Tibetan script allow a stop consonant as a superscript, so there is no way to write the forbidden combination).
• is blocked before དྲ
  • ?with some exceptions, explained as hypercorrection.

The suffix -ས

• is blocked after finals from rows 2, 3, and 5 (palatal, dental, and dental affricates): བ+TAD+ས > བཏད་.
  • NB: In the older orthography, ས is not blocked after consonants of rows 2, 3, and 5, but appears as ད: བ+ZUR+ས > བཟུརད་
• appears as expected after finals from rows 1 and 4 (velars and labials).

The suffix -ད (the present ད་དྲག་)

Certain transitive verbs of classes T1 and T2 take a -ད suffix (the present ད་དྲག་)in the P stem, which manifests in the following ways:

• after a vowel (open syllable), the suffix appears as -ད
• after acute final consonants (those of rows 2, 3, and 5), the -ད suffix is blocked
  • ?retained after acute in old orthography?
• after grave final consonants (those of rows 1 and 4), the -ད suffix appears as -ས

The -ད suffix, in those verbs which take it, also has effects on the stem, regardless of whether it itself is blocked or appears as -ད or -ས:

• The vowel a becomes e: འ+SAM+ད > སེམས་.
• Final ང become ན: འ+LANG+ད > ལེན་, and
  • where this ང > ན change occurs, a u vowel will additionally become i: འ+DZUNG+ད > འཛིན་.

Root adjustment rules

Summary

---

Survey of forms

Prefix changes

• P–TP–TP–T0
• P₁–T0–TP₂–T0. The prefix here may change or may stay the same (NB. exceptionally, sbyin byin sbyin byin can be placed in this class, if we treat the superscript as a prefix)
• P—0—0—0
• (P)-b/m-b/m-(P). Specifically:
  • P—b—b—P, with b replacing (P). Or when verb begins with n-, (P)—m—m—(P).
  • P—b—b—0, with b replacing (P). Or when verb begins with n-, (P)—m—m—(P).
  • 0—b—b—0, with b replacing (P). Or when verb begins with n-, (P)—m—m—(P).
• P—b—g/d—0. P usually = 'a, sometimes g. When verb begins with velar, P—b—d—0.
• 0—b—0—0. Or when verb begins with n-, 0—m—0—0. Only 5 such verbs (klubs, sgur, sdangs, sngam, nod)

---

Root letter changes

It is possible for all four forms of a verb to have the same unchanged root letter, no matter what the root letter is.

15 letters never change. These are ཀ . . ང | . . . ཉ | . . . ན | པ . . མ | . . . . | ཞ ཟ འ ཡ | ར ལ ཤ ས | ཧ .

The remaining 13 possible root letters (. ཁ ག . | ཅ ཆ ཇ . | ཏ ཐ ད . | . ཕ བ . | ཙ ཚ ཛ. ) may undergo changes according to a variety of patterns, or they may be unchanged in all four verb stems, depending on the verb. Most of these are substituted by other consonants of the same row, and so their sequence of substitutes for P, A, F, and I will be indicated by a sequence of column numbers (I, II, III, IV, F=fricative from same column)

Column I consonants: ཅ ཏ ཙ

These consonants follow the pattern I—I—I—II:

• ཅ—ཅ—ཅ—ཆ.
• ཏ—ཏ—ཏ—ཐ.
• ཙ—ཙ—ཙ—ཚ.

Seven verbs undergoing this change take the prefix pattern P—b—g—0 with P=g, i.e. g—b—g—0. These are gcod, gtong, gtum, gtog, gtor, gtsug, and gtsur.

Column II consonants: ཁ ཆ ཐ ཕ ཚ

These only change when prefixed in the P stem with འ, and follow one of four patterns:

• II-I–TI–TII,
• II-II-III-II (three verbs only, all with ཕ)
• II-F-F-II
• II-F-F-F

(where F signifies a fricative of the same column).

Distribution of types by root consonant of present tense

ཁ in present tense root

These follow the pattern II-I–TI–TII:

• ཁ—ཀ—ཀ—ཁ

ཆ in present tense root

These may follow the ཁ pattern, II-I-I-II:

• ཆ—ཅ—ཅ—ཆ,

but they can also follow a pattern with the corresponding fricative, II-F-F-F:

• ཆ—ཤ—ཤ—ཤ

ཐ in present tense root

Like ཁ, these follow the pattern II-I-I-II:

• ཐ—ཏ—ཏ—ཐ

ཕ in present tense root

Three verbs ('phral, 'phri, 'phrog) with this root consonant undergo the pattern II-II-III-II:

• ཕ—ཕ—བ—ཕ

Other consonants of this column (ཁ, ཆ, ཐ, and ཚ)sometimes seem to undergo this pattern (e.g. ཁ—ཁ—ག—ཁ), but in each case the verbs concerned have better attested variants employing one of the other column II patterns.

ཚ in present tense root

Verbs with this root consonant in the P stem can follow either a two-fricative pattern, II-F-F-II

• ཚ—ས—ས—ཚ,

or a three-fricative pattern, II—F—F—F:

• ཚ—ས—ས—ས

or else the same II-I-I-II pattern followed by ཁ, ཆ, and ཐ:

• ཚ—ཙ—ཙ—ཚ.

In this third case, the verb usually takes the prefix pattern (P)-b/m-b/m-(P), which in this case must be འ-b-b-0 or འ-b-b-འ.

Column III consonants

Only undergo changes when the P stem has the prefix འ.

In some of these cases, the prefixes used are predictable.

g in present tense stem

Extremely common class of verbs.
Follows the pattern III-I-II-II:

• g—k—g—kh

The prefixes are certain here, as this group of verbs always follows the prefix pattern P-b-g/d-0, which in this case is realized as འ-b-d-0, giving:

• 'g-bk-dg-kh

j in present tense stem

Most 4-form verbs which have root consonant j in the present tense follow the pattern III-I-F-II:

• j-c-zh-ch

This is probably a subtype of the pattern employed for g and d, with the zh in the F stem arising from an original gj- combination.

Some verbs in j, however, follow the pattern III-F-F-F:

• j-zh-zh-zh

This is analogous to the II-F-F-F pattern which can occur for root consonant ch, i.e. ch-sh-sh-sh.

d in present tense stem

Extremely common class of verbs.
Follows the pattern III-I-II-II:

• d—t—d—th

The prefixes are certain here, as this group of verbs always follows the prefix pattern P-b-g/d-0, which in this case is realized as འ-b-g-0, giving:

• 'd-bt-gd-th

b in present tense stem

Follows the pattern III-II-III-II, and the F stem always has the prefix d:

• b-ph-b-ph, or with prefixes: 'b-?ph-db-?ph (not sure if A and I can have prefixes, or if there's a rule)

This is probably a subtype of the pattern employed for g and d, with the ph in the A stem arising from an original bp- combination.

dz in present tense stem

Most 4-form verbs which have root consonant dz in the present tense follow the pattern III-I-F-II:

• dz-ts-z-tsh

This is probably a subtype of the pattern employed for g and d, with the zh in the F stem arising from an original gdz- combination.

Some verbs in dz, however, follow the pattern III-F-F-F:

• dz-z-z-z

This is analogous to the II-F-F-F pattern which can occur for root consonant tsh, i.e. tsh-s-s-s.

Irregular stems

Hahn (2009) indicates a few verbs which behave in unexpected ways:

• 'jal—bcal—gzhal—'jol
• 'jib—bzhibs—bzhib/gzhib—'jib
• 'co—bzho—bzho—'jo
• Those with (syncretic?) present tense stems, in which the root consonant is unpredictable:
  • ltung—lhung—lhung—lhung
  • ldug—blugs—blug—blugs
  • ldud—blud—blud—blud

---

Vowel changes

A given verb may involve no vowel change at all; that is, the following P, A, F, I sequences are all possible: a-a-a-a, i-i-i-i, u-u-u-u, e-e-e-e, and o-o-o-o.

In fact, a verb with the vowel u in the P stem must have that vowel in all stems.

Verbs with one of the other four vowels (a, i, e, o) in the P stem have one or two (in the case of e) other patterns:

P	A	F	I
a	a	a	o
i	u	u	u
e	e	e	o
e	a	a	o
o	a	a	o

---

Coda changes

The following patterns occur in the P, A, F, and I stems:

• s²—0—0—0, e.g. འདེབས་ བཏབ་ གདབ་ ཐོབ་
• s²—s²—0—s², e.g. འབིགས་ ཕིགས་ དབིག་ ཕིགས་
• d¹—s¹—0—s¹, e.g. བྱེད་ བྱས་ བྱ་ བྱོས་
• 0—s—0—s, e.g. སྨྲ་ སྨྲས་ སྨྲ་ སྨྲོས་; and final -n in the P stem becomes final -ŋ in the AFI stems, giving n—ŋs—ŋ—ŋs, e.g ལེན་ བླངས་ བླང་ ལོངས་
• 0—0—0—s and final -n in the P stem becomes final -ŋ in the AFI stems, giving n—ŋ—ŋ—ŋs, e.g. འབྱིན་ ཕྱུང་ དབྱུང་ ཕྱུངས་
• In older orthography which employs the secondary d, the following pattern also occurs: 0—d²—0—d², as in རྟེན་ བརྟེནད་ བརྟེན་ རྟེནད་

NB: s² means secondary s—this always follows a consonant.

s¹ means primary s—this always follows the syllable's vowel.

(similarly with primary & secondary d)

---