Lushootseed Numerals

Abstract

Summarizes current knowledge of numerals in Lushootseed (Salishan). Published in ICSNL 55 [http://lingpapers.sites.olt.ubc.ca/files/2020/07/01_ICSNL55_Beck_final.pdf]

Full text

Papers for the International Conference on Salish and Neighboring Languages 55.
D. K. E. Reisinger, Hannah Green, Marianne Huijsmans, Gloria Mellesmoen, and Bailey Trotter (eds.).
Vancouver, BC: UBCWPL, 2020.
Lushootseed Numerals*
David Beck
University of Alberta
Abstract: This paper presents a synthesis of what we currently know about numerals and counting
in Lushootseed, gleaned from existing descriptions and texts. Lushootseed has three numeral
series — a plain series for general counting, a reduplicated human series for counting people, and
a temporal-iterative series for counting time and repetitions. Numerals are used both to quantify
nouns and as sentence predicates, and combine with lexical suffixes to form numeral classifier
constructions.
Keywords: Salishan, Lushootseed, numerals
1 Introduction
Aside from lists of forms in dictionaries, passing mention in grammars, and practical lessons in
counting elaborated for teaching, there is very little in the literature on Salishan numerals and
counting — hence, the modest goal of this paper, which is to synthesize what we know about
numerals and numeral phrases in Lushootseed and make it available in a single place. In what
follows, I present whatever can be gleaned from the existing descriptive and pedagogical
grammars (Hess & Hilbert 1976; Hess 1995, 1998, 2006; Cort 1998), a brief technical sketch
(Tweddell 1950), the Lushootseed Dictionary (Bates et al. 1994), extant published texts (Hilbert
& Hess 1977; Beck & Hess 2014, 2015), and a few forms from as-yet unpublished recordings
collected by Vi Hilbert, Thom Hess, and Leon Metcalf (transcribed by Hilbert and Hess) in the
current Lushootseed corpus. What emerges is a surprisingly complete picture, a tribute to the
quality of earlier documentation of the language, but there are nevertheless some unanswered
questions, particularly with regards to the productivity of certain derivational patterns and
ordering restrictions in numeral phrases. It is hoped that as work progresses, we will find answers
to some of these questions, and the search for answers will be advanced by having what we do
know consolidated and synthesized in the pages below. I will begin the discussion by reviewing
the three numeral series — plain, human, and temporal-iterative — and discussing how the units
for digits, decades, and centuries are combined to form complex numerals. Following this, I will
draw on sentences in the corpus to illustrate the uses of numbers and the structure of the numeral
phrase, including the attested use of lexical suffixes as numeral classifiers.
2 Numeral series
Lushootseed has three series of words used for counting. The first of these are the general, plain-
series numerals for counting non-humans given in Table 1.
1
Several of the forms here are
analyzable as consisting of a radical (√) plus the lexical suffix -ačiʔ ‘hand’. In both Northern
Lushootseed (NL) and Southern Lushootseed (SL), this suffix is used to indicate decades (that is,
the interval corresponding to the number of fingers on both hands). The word for 100, sbək’ʷačiʔ,
* Numberless thanks go to Thom Hess, Vi Hilbert, and to the Lushootseed speakers and Elders who gave
their time and their words to make this work possible.
1
Forms in the table are drawn from the full range of sources listed in the references; forms taken from
Tweddell (1950) (Southern Lushootseed forms greater than 20) are given in what is now the standardized
Lushootseed orthography.

2
appears to contain the radical √bək’ʷ ‘all’, suggesting an etymological source in an expression
meaning ‘all hands’. The word for eight, təqačiʔ, is based on the radical √tq ‘closed’, and means
‘closed hands’. According to Hess and Hilbert (1976:28), this is derived from the practice of
counting on one’s fingers by folding them down over the palm, eight being when all the fingers
are closed, leaving only the thumbs. The Southern Lushootseed word for six, dᶻəlačiʔ, is based on
the bound (º) radical º√dᶻəl ‘turned, changed’, which probably derives from the fact that, when
counting on one’s fingers, six requires the use of the fingers on the next hand. Of the two forms
for ‘ten’, the Southern Lushootseed padac appears to be the reflex of the Proto-Salishan *ʔupán-
akis(t) proposed by Kinkade (2002); the origin of the Northern Lushootseed ʔulub is unknown.
The words for 30, 40, 50, and 100 begin with the nominalizing prefix s-.
Table 1: Cardinal numerals — plain series
1
dəč’uʔ
90
x
̌ʷəlačiʔ
2
saliʔ
100
dəč’uʔ sbək’ʷačiʔ
3
ɬixʷ
101
dəč’uʔ sbək’ʷačiʔ ʔi kʷi dč’uʔ
4
buus
102
dəč’uʔ sbək’ʷačiʔ ʔi kʷi saliʔ
5
cəlac
110
dəč’uʔ sbək’ʷačiʔ ʔi kʷi ʔulub (NL)
6
yəlaʔc (NL), dᶻəlačiʔ (SL)
111
dəč’uʔ sbək’ʷačiʔ ʔi kʷi ʔulub ʔi kʷi dč’uʔ (NL)
7
c’ukʷs
120
dəč’uʔ sbək’ʷačiʔ ʔi kʷi saliʔačiʔ
8
təqačiʔ
200
saliʔ sbək’ʷačiʔ
9
x
̌ʷəl
300
ɬixʷ sbək’ʷačiʔ
10
ʔulub (NL), padac (SL)
400
buus sbək’ʷačiʔ
11
ʔulub ʔi kʷi dč’uʔ (NL)*
500
cəlac sbək’ʷačiʔ
12
ʔulub ʔi kʷi saliʔ (NL)
600
yəlaʔc sbək’ʷačiʔ (NL), dᶻəlačiʔ sbək’ʷačiʔ (SL)
20
saliʔačiʔ
700
c’ukʷs sbək’ʷačiʔ
21
saliʔačiʔ ʔi kʷi dč’uʔ
800
təqačiʔ sbək’ʷačiʔ
22
saliʔačiʔ ʔi kʷi saliʔ
900
x
̌̌əl sbək’ʷačiʔ
30
sɬixʷačiʔ
1000
ʔulub sbək’ʷačiʔ (NL), padac sbək’ʷačiʔ (SL)
40
sbuusačiʔ
1001
ʔulub sbək’ʷčiʔ ʔi kʷi dč’uʔ (NL)
50
scəlacačiʔ
1100
ʔulub ʔi kʷi dč’uʔ sbək’ʷačiʔ (NL)†
60
yəlaʔcačiʔ (NL), dᶻəlačiʔačiʔ (SL)
1900
ʔulub ʔi kʷi x
̌ʷəl sbək’ʷačiʔ (NL)
70
c’ukʷsačiʔ
2000
saliʔačiʔ sbək’ʷačiʔ
80
təqačiʔačiʔ
3000
sɬixʷačiʔ sbək’ʷačiʔ
* Literally, ‘ten and one (thing)’. The SL form for eleven is padac yəxʷ kʷi dəč’uʔ (Hess & Hilbert
1976), which would have the same literal gloss.
† Tweddell (1950) reports SL 1,100 as padac sbək’ʷačiʔ yəxʷ kʷi dč’uʔ sbək’ʷačiʔ; however, this
form does not follow the pattern for the other thousand-hundred forms found in the same work and
in Hess and Hilbert (1976).
Numbers that combine decades or centuries with digits such as the Northern Lushootseed
words for 11 (ʔulub ʔi kʷi dč’uʔ), 21 (saliʔačiʔ ʔi kʷi dč’uʔ), or 101 (dəč’uʔ sbək’ʷačiʔ ʔi kʷi
dč’uʔ) use the conjunction ʔi ‘and’ (yəxʷ ‘and’ is used in the Southern Lushootseed equivalents)
and a determiner to connect the digits to the remainder of the expression. The choice of
determiner is governed by the same considerations of specificity, uniqueness, and deixis that
govern determiner choice in other contexts, kʷi ‘remote/hypothetical’ being used as a default
when one is counting in the abstract and not enumerating concrete objects. Orders of centuries are

3
specified by combining words for digits and decades with sbək’ʷačiʔ ‘100’, making the
Lushootseed numerals between 100 and 10,000 the equivalent of those used in spoken English for
numbers (e.g., cəlac sbək’ʷačiʔ = five hundred, ʔulub ʔi kʷi x
̌ʷəl sbək’ʷačiʔ = nineteen hundred),
although in Lushootseed — but not in English — this pattern is extended to the millennia (ʔulub
sbək’ʷačiʔ ‘one thousand’ [lit. ‘ten hundred’]).
2
Lushootseed has a special human series of numerals used specifically for counting people,
given in Table 2. With the exception of the words for one, two, and four people, this series is
formed from the plain-series numerals by Type III (-V1C2- infixal) reduplication.
3
For numerals
between 10 and 20 that combine the first decade with a digit, both numerals are taken from the
human series:
4
(1) ʔululub ʔi diič’uʔ
ʔululub ʔi diič’uʔ
tenHMN and oneHMN
‘eleven people’ (Louise Anderson from Beck & Hess 2015:356, line 56)
In Northern Lushootseed, the human series of numerals is only used up to 20, after which plain-
series numerals are used. In Southern Lushootseed, the complex numerals for 100 and 1000
people are given in Tweddell (1950:72) with the initial numerals in the corresponding human
form, diič’uʔ bəkʷ’ačiʔ ‘100 people’ and padadac bəkʷ’ačiʔ ‘1,000 people’ (lit. ‘ten hundred
people’); the higher thousands such as 2,000 and 5,000 use only the plain cardinal forms (i.e.,
saliʔačiʔ sbək’ʷačiʔ ‘2,000 people’ [lit. ‘twenty-hundred people’] and cəlacačiʔ sbək’ʷačiʔ ‘5,000
people’ [lit. ‘fifty-hundred people’]).
Table 2: Cardinal numerals — human series
1
diič’uʔ
8
təqqačiʔ
2
səsaʔliʔ
9
x
̌ʷələl
3
ɬixʷixʷ
10
ʔululub (NL), padadac (SL)
4
bəbuʔs
11
ʔululub ʔi diič’uʔ (NL)*
5
cələlac
20
salaʔačiʔiʔ (SL)
6
yələlaʔc (NL), dᶻəlalačiʔ (SL)
100
diič’uʔ bəkʷ’ačiʔ (SL)
7
c’ukʷukʷs
1000
padadac bəkʷ’ačiʔ (SL)
*The form padadačiʔ yəxʷ kʷi diič’uʔ is given for SL by Tweddell (1950: 72), who reports
the form padadačiʔ ‘ten’ as being used in compound numbers from 11–19. The form
padadac is used for 10 and as the initial member of compounds based thereon — e.g.,
padadac bəkʷ’ačiʔ ‘1000’.
2
Tweddell (1950:72) also reports the borrowing ta’wowsəd [sic] ‘thousand’.
3
See Anderson (1999) for a survey of reduplicated numeral forms across the Salishan family.
4
The abbreviations used in this paper are as follows: 1, 2, 3 first, second, third person; ADD additive; ADNM
adjunctive nominalizer; ATTN attenuative; CNTRPT centripetal; CNN connective; DAT dative applicative; DC
diminished control; DIST distal; DMA demonstrative adverbial; DSTR distributive; ECS external causative;
FEM feminine; FOC focus; HAB habitual; HMN human classifier; ICS internal causative; INCH inchoative; INT
interrogative; IRR irrealis; MAP middle applicative; NM nominalizer; PASS passive; PAST past tense; PFV
perfective aspect; PL plural; PO possessive; PR preposition; PROG progressive aspect; PROP propriative; PROX
proximal; REFL reflexive; REM remote; SBJ subjunctive; SCONJ sentential conjunction; SG singular; SPEC
specific; STAT stative aspect; SUB subject. Sentences drawn from texts are cited by the name of the speaker
and, where available, a published source.

4
A third series of numerals, given in Table 3, is used for counting time or iterations. The first of
forms in this table, dəč’axʷ ‘once’, is commonly truncated to č’əʔaxʷ. The second form cəbab
‘twice’ is based on a bound suppletive form of the radical √saliʔ ‘two’. It usually appears in
combination with certain lexical suffixes. The remainder of the temporal-iterative numerals are
formed by combining the plain-series numeral with the lexical suffix -aɬ ‘times’. For simple
numerals, this suffix is simply added to the stem; for complex numerals, it appears as a suffix on
the decade or century (or the century if both are present) rather than on the digits. The same
lexical suffix is also used to form the expressions qahaɬ ‘many (qah) times’ and dᶻixʷaɬ ‘first
(dᶻixʷ) time’.
Table 3: Cardinal numerals — temporal-iterative series
5
1
dəč’axʷ
30
sɬixʷačiʔaɬ
2
cəbab
40
sbuusačiʔaɬ
3
ɬixʷaɬ
50
scəlacačiʔaɬ
4
buusaɬ
60
dᶻəlačiʔačiʔaɬ (SL)
5
cəlacaɬ
70
c’ukʷsačiʔaɬ
6
yəlaʔcaɬ (NL), dᶻəlačiʔaɬ (SL)
80
təqačiʔačiʔaɬ
7
c’ukʷaɬ
90
x
̌ʷəlačiʔaɬ
8
təqačiʔaɬ
100
dəč’uʔ sbək’ʷačiʔaɬ
9
x
̌ʷəɬaɬ
101
dəč’uʔ sbək’ʷačiʔaɬ yəxʷ kʷi dč’uʔ
10
ʔulubaɬ (NL), padacaɬ (SL)
200
saliʔ sbək’ʷačiʔaɬ
11
ʔulubaɬ ʔi kʷi dč’uʔ (NL)
700
cukʷs sbək’ʷačiʔaɬ
12
ʔulubaɬ ʔi kʷi saliʔ (NL)
1000
padac sbək’ʷačiʔaɬ (SL)
20
saliʔačiʔaɬ
1001
padac sbək’ʷčiʔaɬ yəxʷ kʷi dč’uʔ (SL)
21
saliʔačiʔaɬ ʔi kʷi dč’uʔ
1100
padac yəxʷ kʷi dč’uʔ sbək’ʷačiʔaɬ (SL)
22
saliʔačiʔaɬ ʔi kʷi saliʔ
2000
saliʔačiʔ sbək’ʷačiʔ
The temporal-iterative numeral series is also the basis of a few complex temporal expressions
derived using the inchoative suffix -il to create verbs indicating the completion of an action a
specified number of times:
(2) x
̌ʷul’ ʔubuusaɬil tiʔiɬ shuyuds …
x
̌ʷul’ ʔu-buus•aɬ-il tiʔiɬ s=huyu-d=s
only PFV-four•times-INCH DIST NM=be.done-ICS=3PO
‘just four times she does that …’
(Mary Sampson Willup from Beck & Hess 2015:289, line 131)
(3) ʔal kʷi sɬixʷaɬils, huy, kʷədabacdubutəxʷ tiʔəʔ ʔəskikəwič
ʔal kʷi s=ɬixʷ•aɬ-il=s huy kʷəd•abac-dxʷ-but=əxʷ
PR REM NM=three•times-INCH=3PO SCONJ take•body-DC-REFL=now
tiʔəʔ ʔəs-C1i-kəwič
DIST STAT-ATTN-hunchbacked
‘on the third time, well, the little hunchbacked one caught his body [on the snags]’
(Agnes James from Beck & Hess 2014:578, line 59)
5
The forms higher than 20 in this table are Snoqualmie-Duwamish forms taken from Tweddell (1950:72)
and given here in standardized transcription; “SL” is used in the table where these forms are predicted to be
different from the Northern Lushootseed forms, which are so far unattested.

5
It may be, however, that this particular type of derivation is not limited to temporal-iterative
numerals as there are two examples from texts, given in (4) and (5), of inchoative verbs formed
with a numeral and a different lexical suffix, -gʷiɬ ‘canoe’:
(4) ləcəbagʷiɬil tiʔəʔ ʔəsx
̌ək’ʷtxʷ əlgʷəʔ ƛ’əlayʔ
lə=cəb•a•gʷiɬ-il tiʔəʔ ʔəs-x
̌ək’ʷ-txʷ əlgʷəʔ ƛ’əlayʔ
PROG=two•CNN•canoe-INCH PROX STAT-overturned-ECS PL canoe
‘it was coming to be that they had two canoes overturned’
(5) ləɬixʷalgʷiɬil tiʔəʔ tasx
̌ək’ʷtxʷ həlgʷəʔ ƛ’əlayʔ dəxʷəsax
̌ʷəbabac
lə=ɬixʷ•al•gʷiɬ-il tiʔəʔ tu=ʔas-x
̌ək’ʷ-txʷ həlgʷəʔ ƛ’əlayʔ
PROG=three•CNN•canoe-INCH PROX PAST=STAT-overturned-ECS PL canoe
dəxʷ=lə=sax
̌ʷəb•abac
ADNM=PROG=jump•body
‘it was coming to be they had three canoes overturned that they were jumping over’
(Harry Moses from Hilbert & Hess 1977:15)
This suggests that the numeral+lexical suffix+inchoative expression may be more productive than
its frequency in the corpus indicates, but this will have to remain an open question until further
textual analysis brings more examples to light.
Otherwise, numerals in all three series are surprisingly inert in morphological terms. Beyond
expressions of the type shown above, numerals seem only to appear as stems in two related sets
of compound words formed from the temporal-iterative numerals and the lexical suffix -dat ‘day’,
shown in Table 4. These seem likely to be recent formations introduced along with the European
calendar. The fact that there are a maximum of four forms in each set follows from the naming
practices for the days of the week, the remainder of which are č’itabac ‘Saturday’ (from √č’it
‘near’ + -abac ‘body’), x
̌aʔx
̌aʔaɬdat ‘Sunday’ (√x
̌aʔx
̌aʔ ‘sacred, taboo’), and bəlx
̌ʷəɬdat ‘Monday’
(√bəlx
̌ʷ ‘pass by, come after’). Furthermore, it seems possible that the words in the left-hand
column of Table 4 are backformations from the (more morphologically complex) words in the
righthand column, given the absence of potential forms for expressing periods of time in days that
do not correspond to the Lushootseed names for days of the week that happen to be based on
numerals. It should be noted, however, that there are two verbs — ɬixʷəɬdacut ‘become three
days’ and buusaɬdalicut ‘become four days’ (each attested only once in the corpus) — which
appear to be derived from sɬixʷəɬdat ‘three days, third day’ and buusəɬdat ‘four days, fourth day’,
respectively. These two verbs are formed through a rather non-transparent use of reflexive
morphology (-cut consisting of the event-internal causative -t and the reflexive marker -sut), and
it remains to be seen whether they are representative of a more extensive set of older forms or if
they are recent coinages.
Table 4: Compound numeral expressions for days and days of the week
cəbdat ‘two days, second day’
scəbdatil ‘Tuesday’
sɬixʷəɬdat ‘three days, third day’*
sɬixʷəɬdatil ‘Wednesday’
buusəɬdat ‘four days, fourth day’
buusəɬdatil ‘Thursday’ (NL)
—
scəlacəɬdatil ‘Friday’ (NL)†
*This form is also glossed as ‘Wednesday’ in Bates et al. (1994:147).
†Bates et al. (1994:45) records this form as scəlacəɬdat(il), indicating it is also
attested without the inchoative suffix but with the same gloss.

6
3 Numerals and numeral phrases
All three series of numerals are used both as adnominal quantifiers and as clausal predicates. The
former use is illustrated in (6)–(8):
(6) hay gʷəl, tusulayitəbəxʷ ʔə tiʔiɬ saliʔ sq’ʷiq’ʷəlaƛ’əd
hay gʷəl tu=sula-yi-t-b=əxʷ ʔə tiʔiɬ saliʔ sC1i-q’ʷəlaƛ’əd
SCONJ SCONJ PAST=centred-DAT-ICS-PASS=now PR DIST two ATTN-berry
‘and then he set before him two little berries’
(Edward Sam from Beck & Hess 2014:536, line 21)
(7) xʷuʔələʔ ʔəsʔəx
̌id cəlac sləx
̌il kʷədiʔ tusaxʷəbtubs
xʷuʔələʔ ʔəs-ʔəx
̌id cəlac sləx
̌il kʷədiʔ tu=s=saxʷəb-txʷ-b=s
maybe STAT-transpire five day REM.DMA PAST=NM=run-ECS-PASS=3PO
‘maybe it was five days (since) they had been run off with (i.e., kidnapped)’
(Martha Lamont from Beck & Hess 2014:108, line 265)
(8) gʷəl ʔahəxʷ tiʔəʔ səsaliʔ sɬəɬadəyʔ ləqaladibid
gʷəl ʔah=əxʷ tiʔəʔ səsaliʔ sC1ə-ɬadəyʔ ləq•al•adiʔ-bi-d
SCONJ be.there=now PROX twoHMN PL-woman hear•CNN•ear-MAP-ICS
‘and there are two women who overhear it’
(Harry Moses from Beck & Hess 2015:298, line 27)
Numerals are most frequently used with the singular form of nouns, as in (6) and (7), although the
plural form is an option, particularly when referring to numbers of people, as in (8). Numerals
may be used in argument phrases, as in (6) and (8), as well as in predicate complements (7).
The numeral ‘one’, dəč’uʔ, can be used to convey a sense of specificity or particularity:
(9) gʷəl huy bəʔəy’ədxʷəxʷ tiʔəʔ qa tiʔiɬ sʔuladxʷ ʔal tiʔiɬ cədiɬ dəč’uʔ stuləkʷ
gʷəl huy bə=ʔəy’-dxʷ=əxʷ tiʔəʔ qa tiʔiɬ sʔuladxʷ ʔal tiʔiɬ
SCONJ SCONJ ADD=found-DC=now PROX many DIST salmon at DIST
cədiɬ dəč’uʔ stuləkʷ
s/he one river
‘and then he found a lot of salmon in this one river’
(Martha Lamont from Beck & Hess 2014:202, line 23)
(10) gʷəl diɬ xʷuʔələʔ tušac’s tiʔiɬ dəč’uʔ syəyəhub
gʷəl diɬ xʷuʔələʔ tu=s=šac’=s tiʔiɬ dəč’uʔ syəyəhub
SCONJ FOC maybe PAST=NM=end=3PO PROX one story
‘and I guess that is the end of this one traditional story’
(Martha Lamont from Beck & Hess 2014:236, line 279)
(11) ʔaləxʷ tiʔəʔ sʔahilsəxʷ tiʔəʔ sɬčils dxʷʔal tiʔəʔ dəč’uʔ swətixʷtəd
ʔal=əxʷ tiʔəʔ s=ʔah-il=s=əxʷ tiʔəʔ s=ɬč-il=s
at=now PROX NM=be.there-INCH=3PO=now PROX NM=arrive-INCH=3PO
dxʷ-ʔal tiʔəʔ dəč’uʔ swətixʷtəd
CNTRPT-at PROX one tree
‘when (she) got to this one tree’ (Martin Sampson from Beck & Hess 2015:379, line 41)

7
(12) gʷəl ƛ’ubəkʷədad tiʔiɬ bədəč’uʔ qʷɬayʔ
gʷəl ƛ’u=bə=kʷəda-d tiʔiɬ bə=dəč’uʔ qʷɬayʔ
SCONJ HAB=ADD=taken-ECS DIST ADD=one stick
‘then she would take another one of her scratching sticks’
(Alice Williams from Beck & Hess 2015:424, line 123)
This use of the numeral is not textually infrequent and finds a very close parallel in its literal
English gloss.
Like other adnominal modifiers, numerals have some flexibility as to whether they precede
the noun they modify, as in (6) and (8) above, or follow it, as in (13) and (14):
(13) təɬ čəd ʔuʔəy’dxʷ tiʔiɬ dsqa cəxʷsqatəd səsaʔliʔ ʔal tudiʔ
təɬ čəd ʔu-ʔəy’-dxʷ tiʔiɬ d-sqa
truly 1SG.SUB PFV-found-DC DIST 1SG.PO-older.sibling
d=dəxʷ=sqatəd səsaʔliʔ ʔal tudiʔ
1SG.PO=ADNM=older.siblings twoHMN at DIST.DMA
‘I truly did find my older brothers, my two older brothers over there’
(Martha Lamont from Beck & Hess 2014:172, line 739)
(14) sx
̌aʔhus tsiʔəʔ čəgʷas diič’uʔ
sx
̌aʔhus tsiʔəʔ čəgʷas diič’uʔ
sawbill PROX:FEM wife oneHMN
‘one wife [was] Sawbill’ (Martha Lamont from Beck & Hess 2014:447, line 5)
Post-nominal position for numerals is rare (these are the only two examples in the current corpus)
and it is not clear what conditions may apply to this ordering.
Also like other modifiers, numerals can be combined with additional adnominal elements in a
single noun phrase:
(15) saʔ saliʔ sq’ʷəlaɬəd
saʔ saliʔ sq’ʷəlaɬəd
bad two berry
‘[there were] two measly berries’ (Edward Sam from Beck & Hess 2014:537, line 25)
(16) dəč’uʔ haʔɬ syəyəhub
dəč’uʔ haʔɬ syəyəhub
one good story
‘[it is] a good story’ (Harry Moses from Hilbert & Hess 1977:32)
The relative ordering of the numeral with respect to other adnominal modifiers appears to be
relatively free, determined by considerations of style or communicative structure.
Numerals are often found as the heads of anaphoric nominal expressions:
(17) kʷədad tiʔəʔ dəč’uʔ
kʷəda-d tiʔəʔ dəč’uʔ
taken-ICS PROX one
‘he took one (berry)’ (Edward Sam from Beck & Hess 2014:537, line 28)

8
(18) gʷəl bəʔəsgʷədil tsiʔiɬ dəč’uʔ ʔal kʷi xʷqʷəq’ʷus
gʷəl bə=ʔəs-gʷəd-il tsiʔiɬ dəč’uʔ ʔal kʷi xʷqʷəq’ʷus
SCONJ ADD=STAT-down-INCH DIST.FEM one at REM cliff
‘and one (sister) was sitting on the cliff’ [Julia Siddle Basket Ogress, line 74]
These expressions only occur in contexts where the identity of the item(s) being counted is
recoverable from discourse. Numerals also head NPs in constructions such as that in (19):
(19) gʷəl ʔəbsqʷəbqʷəbayʔ ʔə tiʔəʔ bəsaliʔ
gʷəl ʔəs-bəs-sC1V1C2-qʷəbayʔ ʔə tiʔəʔ bə=saliʔ
SCONJ STAT-PROP-DSTR-dog PR PROX ADD=two
‘and (they) had two dogs, too’ (Martha Lamont from Beck & Hess 2014:35, line 24)
This is a usual way of expressing the notion ‘X has a certain number of Y’ when the fact of
possession, rather than the number possessed, is the focus of the utterance.
Perhaps even more frequently than they are found as part of NPs, numerals occur in predicate
position as the heads of clauses:
(20) saliʔ kʷi ɬuʔəƛ’txʷ čəxʷ č’ƛ’aʔ
saliʔ kʷi ɬu=ʔəƛ’-txʷ čəxʷ č’ƛ’aʔ
two REM IRR=come-ECS 2SG.SUB stone
‘the stones that you will bring [will be] two’
(Alice Williams from Beck & Hess 2015:418, line 85)
(21) saliʔ tiʔiɬ ʔəskʷədad
saliʔ tiʔiɬ ʔəs-kʷəda-d
two DIST STAT-taken-ICS
‘what he is holding [are] two (halibut)’
(Martha Lamont from Beck & Hess 2014:116, line 329)
(22) ɬixʷixʷ tiʔəʔ caadiɬ təlixʷ suq’ʷaʔ
ɬixʷixʷ tiʔəʔ caadiɬ təlixʷ suq’ʷaʔ
threeHMN PROX they blood.brother younger.sibling
‘these full-blood brothers [were] three’
(Martha Lamont from Beck & Hess 2014:226, line 204)
Like other non-verbal predicates, numerals in this context take clitics for mood and tense:
(23) gʷəl gʷətusəsaliʔ əlgʷəʔ
gʷəl gʷə=tu=səsaliʔ əlgʷəʔ
SCONJ SBJ=PAST=twoHMN PL
‘and there would have been two of them’
(Martha Lamont from Beck & Hess 2014:281, line 151)
There is also one example in the corpus of a numeral taking an aspectual prefix:

9
(24) ʔəsbuus kʷi tuhuyud əlgʷəʔ
ʔəs-buus kʷi tu=huyu-d əlgʷəʔ
STAT-four REM PAST=make-ICS PL
‘what they made [was] four [moccasins]’
(Dora Solomon from Beck & Hess 2015:258, line 373)
However, the fact that there is a single occurrence of an aspect marker on a numeral in the corpus
suggests that this is a rather marginal construction. That it occurs on the numeral buus ‘four’ may
not be coincidental, given that four is a culturally important numeral in Northern Lushootseed
(Hess 1995),
6
suggesting perhaps that buus might be a prime candidate for conversion to a verb
expressing the notion of making/doing something in fours.
Like any other predicate, numerals can head yes/no interrogatives and can take ordinary
matrix-clause subject markers:
(25) səsaliʔ ʔu
səsaliʔ ʔu
twoHMN INT
‘were there two (children)?’ (Martha Lamont from Beck & Hess 2014:77, line 37)
(26) ɬixʷačiʔəxʷ čəd ʔi kʷi yəlaʔc
ɬixʷačiʔ=əxʷ čəd ʔi kʷi yəlaʔc
thirty=now 1SG.SUB and REM six
‘I’m thirty-six now’ (Bates et al. 1994:277)
As shown in (26), complex numerals in predicate position are treated like other multi-word
predicate phrases in terms of the placement of subject-markers and other sentence-second clitics.
Numerals in both argument and predicate phrases may be combined with lexical suffixes
acting as numeral classifiers:
(27) … tusaxʷəbabacəd tiʔəʔ buusalgʷiɬ ƛ’əlayʔ
tu=saxʷəb•abac-əd tiʔəʔ buus•al•gʷiɬ ƛ’əlayʔ
PAST=run•body-ICS PROX four•CNN•canoe shovel.nose.canoe
‘… [they] jumped over the four shovel-nosed canoes’
(Harry Moses from Hilbert & Hess 1977:16)
(28) cəbagʷiɬəxʷ kʷi ɬudəxʷsaxʷəbabacəds əlgʷəʔ
cəb•a•gʷiɬ=əxʷ kʷi ɬu=dəxʷ=saxʷəb•abac-əd=s əlgʷəʔ
two•CNN•canoe=now REM IRR=ADNM=jump•body-ICS=3PO PL
‘now what they were jumping over [was] two canoes’
(Harry Moses from Hilbert & Hess 1977:15)
6
The number four is also culturally important in a number of neighbouring but unrelated languages, and the
form *moos proposed by Edward Sapir is a plausible early areal borrowing (Lushootseed having changed
*m to /b/ in the mid-nineteenth century). Among the Southern Lushootseed, the numeral is five (Hess
1995).

10
(29) č’əʔilc ti dtalə
č’əʔ•ilc ti d-talə
one•round SPEC 1SG.PO-dollar
‘I have one dollar’ (Hess & Hilbert 1976:I, 68)
(30) ʔəsɬixʷulč ti dsʔax
̌ʷuʔ
ʔəs-ɬixʷ•ulč ti d-sʔax
̌ʷuʔ
STAT-three•container SPEC 1SG.PO-clam
‘I have three clams’ (Hess 1995:20)
The choice of suffix is largely semantic: lexical suffixes with very concrete, specific meanings are
used in counting those objects which they designate, as in (27) and (28), while others are used
when counting objects that fit into the general class (in either shape or function) of things
expressed by the suffix, as in (29) and (30). Also illustrated by (28) and (30) are the alternate
combining forms for the numerals one and two, č’əʔ- and ‘one’ cəb- ‘two’, mentioned above in
the context of the temporal-iterative series of numerals in Table 3. These forms seem to be in free
variation with the regular stems, dəč’uʔ ‘one’ and saliʔ ‘two’ in enumerative constructions.
4 Some final thoughts
Although the coverage of numerals and numeral phrases in the existing descriptive materials on
Lushootseed is quite good, making it possible to know the way that large complex numbers were
formed and how numerals functioned in syntax, the fact remains that the analyzed corpus is
relatively small and there are still some gaps in our knowledge of how numbers work in
Lushootseed. One particularly striking thing I noted from the corpus (the 26 texts in Beck & Hess
2014, 2015 and a half dozen others we have digitized so far) is dearth of numeral classifier
constructions used in counting/enumeration. Given the richness of the system described for the
Halkomelem languages (Gerdts et al. 2002; Shaw et al. 2002), which use about 30 different
lexical suffixes for counting objects, the paucity of examples from the texts (there are only four,
all using -gʷiɬ ‘canoe’) is surprising, as is the lack of diversity in the suffixes themselves (there
are only seven other suffixes attested with numerals in the descriptive literature). It is hard to
know if this an accidental gap, if it is because the system in Lushootseed is simply less elaborate
than that of many of its congener languages, or if it is the result of the kind of language attrition
Shaw et al. (2002) observe in hən’q’əmin’əm’ — although the last option seems unlikely given
that the analyzed corpus contains speech from very fluent, Lushootseed-dominant Elders. If this
were the case, it would have had to have been a long-standing process, more contact-induced
language change than language loss.
The fact that there are marked differences in numeral forms between the two main dialectal
divisions of Lushootseed is also interesting, and hints at early sociolinguistic patterns and the
possible borrowing of forms from neighbouring languages. Further comparison with the numeral
systems in other languages of the broader family is also in order, and it is to be hoped that in the
coming years, a little more attention can be paid to Salishan numerals. They have a complex
history, with a surprising heterogeneity of forms and etymologies, and numerals in the three
“Mosan” language families of the Central Northwest Coast show the effects of intense language
contact and borrowing (see, for example, Kinkade 2002). Their continued study is sure to provide
deeper insight into the early Northwest Coast linguistic landscape.

11
References
Anderson, Gregory D. S. 1999. Reduplicated numerals in Salish. International Journal of
American Linguistics 65, 407–481.
Bates, Dawn, Thomas M. Hess, and Vi Taqwšəblu Hilbert. 1994. Lushootseed Dictionary.
Seattle: University of Washington Press.
Beck, David, and Thom Hess. 2014. Tellings from Our Elders: Lushootseed syəyəhub. Volume 1,
Snohomish texts. Vancouver: UBC Press.
Beck, David, and Thom Hess. 2015. Tellings from Our Elders: Lushootseed syəyəhub. Volume 2,
Tales from the Skagit Valley. Vancouver: UBC Press.
Cort, David. 1998. Teaching Lushootseed math. In: Papers for the 33rd International Conference
on Salish and Neighboring Languages, 85–90.
[http://lingpapers.sites.olt.ubc.ca/files/2018/03/1998_Cort.pdf]
Gerdts, Donna B., Mercedes Q. Hinkson, and Thomas E. Hukari. 2002. Numeral Classifiers in
Halkomelem. In: Papers for the 37th International Conference on Salish and Neighboring
Languages, UBCWPL 9, 147–180.
[http://lingpapers.sites.olt.ubc.ca/files/2018/03/2002_Gerdts_Hinkson_Hukari.pdf]
Hess, Thom. 1995. Lushootseed reader with introductory grammar, Volume I: Four Stories from
Edward Sam. Missoula: University of Montana Occasional Papers in Linguistics.
Hess, Thom. 1998. Lushootseed reader with intermediate grammar, Volume II: Four Stories from
Martha Lamont. Missoula: University of Montana Occasional Papers in Linguistics.
Hess, Thom. 2006. Lushootseed reader with English Translations, Volume III: Four More Stories
from Martha Lamont. Missoula: University of Montana Occasional Papers in Linguistics.
Hess, Thomas M., and Vi Taqʷšəblu Hilbert. 1976. Lushootseed: An introduction, Books 1 and 2.
Seattle: American Indian Studies, University of Washington.
Hilbert, Vi Taqʷšəblu, and Thomas M. Hess. 1977. Lushootseed. In Barry F. Carlson (ed.),
Northwest Coast texts: Stealing light, 4–32. Chicago: University of Chicago Press.
Kinkade, M. Dale. 2002. Salish numerals in “Old” Nitinaht, In: Papers for the 37th International
Conference on Salish and Neighboring Languages, UBCWPL 9.
[http://lingpapers.sites.olt.ubc.ca/files/2018/03/2002_Kinkade.pdf]
Shaw, Patricia A., Susan J. Blake, and Jill Campbell. 2002. Numerals and lexical suffixes in
hən
̓q
̓əmin
̓əm
̓, In: Papers for the 37th International Conference on Salish and Neighboring
Languages, UBCWPL 9.
[http://lingpapers.sites.olt.ubc.ca/files/2018/03/2002_Shaw_Blake_Campbell.pdf]
Tweddell, Colin Ellidge. 1950. The Snoqualmie-Duwamish dialects of Puget Sound Salish: An
outline of phonemics and morphology. University of Washington Publications in
Anthropology 12. Seattle, University of Washington Press.