Edinburgh Handbook of

Evaluative Morphology
Edited by Nicola Grandi and Livia Kortvelyessy

EDINBURGH
University Press

20.3 Lule

Raoul Zamponi and Willem J. de Reuse

1 Introduction
Lule was once spoken by a semi-nomadic people known under the same name who lived
in the Gran Chaco between the Pilcomayo river and the Andean foothills of north-western
Argentina. It is a dialect of an extinct language, also called Lule (or, less commonly,
Lule-Tonocote), also spoken in that area by other groups: the Ysistine, the Toquistine, the
Oristine and, probably, the Tonocote. This last group, part of the sedentary indigenous
population of the present-day provinces of Tucuman and Santiago del Estero, is treated
as linguistically separate from the Lule by early colonial sources such as the Relaciones
geognijicas de Indias (Jimenez de la Espada 1965, I: 390-6; II: 78-85). They are, however,
connected linguistically to the Lule by the Sardinian Jesuit missionary Antonio Maccioni
(1688-1753) in his grammar and vocabulary of Lule (proper) as shown in the title of his
work: Arte y vocabulario de la lengua lule y tonocote (Machoni 1732). It is likely that the
Tonocote, who suffered constant attacks from the Lule, had shifted to the Lule language
by Maccioni's time.
Mentions of Lule and its co-dialects disappear from the historical record after the
expulsion of the Jesuits in 1776 (Viegas Barros 2001, 14). At the time of renewed interest
in the indigenous languages of Argentina, under the impetus of scholars such as Lafone
y Quevedo (1894), Calandrelli (1896) and Cabrera (1910), it became clear that no Lule
speakers could be found. We may surmise that the Lule language lost its last speakers
sometime in the course of the early nineteenth century.

The external relationships of the Lule language are controversial. The most widely
accepted affiliation is with the moribund Vilela of the Bermejo river basin further east1 to
form the Lule-Vilela family. The possibility cannot be excluded, however, that the shared
vocabulary of the two languages (Viegas Barros 2001) is entirely due to contact, rather than
common ancestry~ and that they are geographically close isolates (Zamponi 2008, li- lv).
More work on the precise historical relationship between Lule and Vilela is therefore a
strong desideratum.
The above-mentioned work by Maccioni, first published in Madrid in 1732, is our
primary source for the Lule language.2 The picture ofLule that emerges from Maccioni's
Arte is that of an agglutinating language with a rather simple and basically suffixal morphology. The language appears to have open classes of nouns, verbs (which includes most
of the words that are semantically adjectives) and adverbs, and closed classes of pronouns,
demonstratives, postpositions and particles (including discourse markers and interjections). Nominal inflectional categories include number (singular vs. plural) and possession.

617

Lule

Verbal inflectional categories include person (of the subject), number, tense and mood. The
derivational morphology is somewhat richer. The language has a set of 'transpositional'
affixes with a specific word-class-changing function and a wider set of 'semantically rich'
affixes (with a concrete meaning) including nine instrumental prefixes (which is quite
unusual for the Chaco area, and actually for South America as a whole) and a set of morphemes expressing evaluative values.

2 The system of evaluative morphology

Eight evaluative suffixes (or morphemes that we may consider as such3) are attested in Lule,
as well as a polysemic reduplicative process which expresses, among other things, attenuation and speed. Most of these suffixes and the reduplicative process are not restricted to a
specific word class, being attested with verbs and nouns, and sometimes also with adverbs.
In the following subsectjons, each of the affixes and the reduplicative process will be
examined as far as semantics and uses are concerned. The forms and the examples are
drawn from Maccioni 's Arte, and will be presented in a tentative phonemic transcription
based on the analysis ofMaccioni's orthography of Lule proposed in Zamponi (2008). 4
2.1 -eks 'celerative'

Used with activity verbs denoting activities that can be carried out at variable speed, the
suffix -eks signals (just like the above-mentioned reduplicative process; see Section 2.9.
and also compare examples (1) and (51)) that a given activity is realised quickly:
(1)

ne-eks-

(2)

come-CBL
'come quickly' (D: 132)
(3)

is-eks-

urinate-CEL
'urinate quickly' (D: 100)

apes-eks-

go.down-CBL
'go down quickly' (0:24)
2. 2 -etsi 'ameliorative '

The suffix -etsi, a grammaticalisation of the verb etsi- 'be good, healthy', appears in the
Arte with two state verbs. It conveys an ameliorative meaning:
(4)

un-etsi-

( 5)

taste-AML
'have a good taste' (D: 118)

akem-etsi

give.off.a.smell-AML
'give off a good smell' (D: 114-15)

It also occurs with state and activity verbs, producing state verbs with the meaning 'be
good to v' or 'worthy ofv-ing'. Interestingly, in these forms, -etsi does not appear adjacent

to the stem of the base verb, but on the right periphery, after the (inflectional) subject marker:
(6)

kai-p-etsi

eat-3SG-AML
'it is good to eat' (0:54)

(7)

amaitsi-p-etsi

love-3SG-AML
'(it is) worthy of love, lovable' (0:70)

Raoul Zamponi and Willem J. de Reuse

618

(8)

unuk-p-etsi
(9)
loathe-3SG-AML
'it is worthy of hate, contemptible'

yet-ip-etsi
dance-3sG-AML
'it is a good time to dance' (0:55)

(0:70)

The stem with which the suffix ~etsi occurs may also include a nominalising reduplicative
process (see also Section 2.9. for more on reduplication), as seen in (1 0) and (11 ):
(1 0)

titi-p-etsi
doNMLS-3SG-AML
'it is good to do' (0:55)

(11)

eutita ka[ka]i-petsi
now
eat[NMLS)-3SG-AML
'now is a good time to eat' (0:55)

Followed by the intensive suffix -ike (Section 2.6.), -etsi is also used with nominal
stems. Added to the noun ekis 'dream', the suffix sequence of -etsi plus -ike indicates that
the referent of the noun is good:
(12)

ekisetsike lek.is-etsi-ikel
dream-AML-INT
'good dream' (D: 123)

The same sequence added to the noun il?e 'elder sister' indicates that the referent of the
noun is a blood relative:
(13)

il?etsikep lil?e-etsi-ike-pj
elder.sister-AML-INT-3SG
'his or her elder sister (by blood)' (0:93)

2.3 -eyu 'pejorative '

Contrasting with -etsi, the suffix -eyu is pejorative, and is a granunaticalisation of the verb
eyu- 'become ill'. It is attested with two state verbs (14)-(15) and one noun (16):
(15)

(14)

un-eyutaste-PBJ
'have a bad taste' (D:13)

(16)

ekis-eyu
dream-PEJ
'nightmare, bad dream' (D: 123)

akem-eyugive.off.a.smell-PEJ
'give off a bad smell' (D:99)

2.4 -(i)pan 'derogatory'

The suffix -pan (postvocalic variant) or -ipan (postconsonantal variant) has a derogatory
connotation and is attested with nouns referring to human beings only: 5
( 17)

!cumue-pan
woman-DER
'prostitute' (D: 112)

Lule

619

It is also attested with two verbal stems, producing two tenns for 'slave':

(18)

met-ipan
takeDER
'slave' (D:36)

( 19)

wenek-ipan
buy-DER
'bought slave' (D:36)

2.5 -ikeps 'augmentative'

The suffix -ikeps is augmentative when used with nouns denoting inanimate referents:
(20)

epuwkwe-ikeps
lightning-son-Auo
'flood' (epukwe 'rain, shower') (D:55)

With a kinship tenn, it indicates that the referent of the noun is a blood relative:
(21)

pepe-ikeps
fatherRDP-AUG
'my father (by blood)' (0:92)

(22)

ya?a-ikeps-ts
brother-AUG-! sg
'my brother (by blood)' (G:93)

It is also used with state verbs. In this case, it has an intensifying meaning:

(23)

etsi-ikeps- letsi-ikeps-1
be.good-Auo
'be very good' (D:94)

(25)

ewikeps-lewi-ikeps-1
be.a.lot-AUG
'be very much' (D:94)

(24) leteps-ikepsbe.damp-AUG
'be very damp' (D:78)
(26)

ap 'u-ikepsbe.stuffed-AUG
'be very stuffed' (D:72)

Finally, -ikeps also appears to intensify the adverb mia 'only' producing a verb meaning
'be just only'. The 3so present indicative fonn of this verb, in tum, also has an adverbial
use attested by (27). This verb can also occur as an adverb with the adverbialising suffix
-le, as in (28):
(27)

mia-ikeps-p
only-AUG-3SG
'just only' (D: 124)

2. 6 -ike

(28)

mia-ikeps-le
only-AUG-ADVLS
'truly' (0:86)

~intensive '

The suffix -ike occurs with state verbs with the effect of intensifying their meaning. The
difference between -ike and -ikeps is not always clear. Pairs of stems attested with both
-ikeps and -ike are in examples (23)-(31), (25)-{35) and (28)-(40):
(29)

waleks-ike(30)
know-rNT
'know a lot, be very wise' (G:63)

amaitsike-lamaitsi-ike-1
love-INT
'desire' (D:53)

620

Raoul Zamponi and Willem J. de Reuse

(31)

etsike- letsi-i.ke-1
be.good-INT
'be very good' (0:94)

(32)

lokots-ikebe.soft-INT
'be very soft (to the touch)' (0:27)

(33)

yeun-ikebe.happy-INT
'be very happy' (G:86)

(34)

kelots-ikebe.clear-INT
'be clean' (0:9)

(35)

ewike-jewi~ike-1

(36)

wele-ikebe.second-INT
'be the last, be final' (D: 108)

be.a.lot-!NT
'be very much' (0:94)

The fonns in (37)-(38) are 3sG indicative present fonns, used adverbially, of verbs derived
by lokots- 'be mild' by means of -ike and of the above-mentioned reduplicative process,
here expressing attenuation (see Section 2.9). Of these forms, that in (38) is unique in that
it contains this reduplicative process twice, once on lokots- and once on -ike. The combined
semantic import of -ike and of attenuative reduplication, both in (37) and in (38), cannot
be derived from Maccioni 's glossing, but whatever it might turn out to be, it is definitely

expresstve:
(37)

loko[ko]ts-ike-p
be.mild[A:IT)-INT-3SG
'mildly' (G:86)

(38)

loko[ko]ts-ike[ke}-p
be.mild[ATI]-INT(ATT]- 3SG
'mildly' (G:86)

The corpus also contains a kinship term noun formed with -ike. Like -ikeps in (21) and
(22), -ike indicates that the referent of the stem noun is a blood relative:
(39)

melu-ike-p
younger.sister-INT-3sG
'her younger sister (of a woman, by blood)' (G:93)

In addition, like -ikeps in (27) and (28), the suffix -ike can intensify the adverb mia 'only'
resulting in a verb whose 3sG indicative present fonn is also used adverbially (40):

(40)

mia-ike-p
only-INT-3SG
'certainly' (D:47)

Finally, (41) is a state verb derived by means of -ike from a stem pili(-) that may be the
noun pili 'wing; feather' or the verb pili- 'fly':
(41)

pilike- IPili-ike-1
wing/feather/fly-INT
'be light (in weight)' (D:83)

Lute

621

2. 7 -ewi 'intensive'
A further intensive suffix is -ewi, a grammaticalisation of the verb ewi- 'be many'. It is

attested in the following state verb:

(42)

pei-ewibe.wide-INT
'be very wide' (0:14)

2.8 -ke 'approximative'

The approximative suffix -ke means 'about, more or less'. It is attested with two adverbs:
(43) and (44):
(43)

eutita-ke
now-APPROX
'a little time ago' (D:8)

(44)

ustse-ke=ma
how-APPROX-INTG
'how, approximately?',
'how, more or less?' (D:40)

2. 9 Reduplication with attenuative, celerative and other meanings

There exists a reduplicative process which, like the reduplicative nominaliser mentioned
in connection with examples (10)-(11) above, copies the (second) consonant of the onset
and the (first) vowel of the nucleus of syllable to which it is joined. However, unlike the
reduplicative nominaliser, it is placed after the first syllable of the polysyllabic stems with
which it occurs. It can be observed in our corpus as a suffix CV after monosyllabic stems
composed of a consonant and a vowel ((47), (48) and (51)) and bisyllabic stems ending
in a CV syllable ((56) and (57b)); as an infix [CV} within stems consisting in a single
syllable provided with onset and coda (45), or whose nucleus coincides with a diphthong
(49); or within stems formed of at least three syllables the second of which is CV ((58)(60)); and within other polysyllabic stems following a syllable with an onset ((46), (50)
and (52)-(55)).
This reduplication is a polysemic process which has five evaluative meanings connected
to the aspectual classes of the base verbs:

attenuative, when applied to state verbs (45)-{48) and activity verbs denoting
events which can be carried out with more or less intensity (49)-(50):

(45)

lu[lu]psbe.hot[xrr]
'be lukewarm'

(47)

p 'op 'obe. wbite/be. dry ATT

(46)

ake[ke]lbe.hot[KIT]
'be lukewarm'

'be whitish, be almost dry' (D:5, 76, 82)

(48)

sisi~

for. there. to. beATr

'be there in a small measure'
(0:61, 73)

Raoul Zamponi and Willem J. de Reuse

622

(49)

(51)

ka[ka)ieat[ATT]
'eat a little' (0:65)

(50)

olo[lo]mspeak[ATT]
'speak a little' (0:65)

celerative~

when applied to activity verbs referring to activities that can be carried

out with more or less speed:
nenecomeCEL
'come quickly' (D: 132)

(52)

yep-ku[ku]iIP-move[CBL]
'move quickly' (D:94)

iterative (or multiplicative), with verbs of achievement:

(53)

oko[kojts(54)
lean. out[ITER]
'lean out many times' (D:21)

tako[ko]tsembrace[ITER]
'embrace many times' (D:2)

incessative, with activity verbs denoting events that can be carried out for an indeterminate length of time:
(55)

takyu[yu]?ulook.for[rNCEss]
'look for incessantly' (D:29)

evolutive, with, apparently, incrementative verbs:

(56)

watotorecover/liveEVOL
'convalesce' (D:41)

The various meanings expressed by this reduplicative process can be regarded as

fonning a continuum. This continuum is attenuative at one end and celerative at the
other, and can be represented as follows: 'attenuative' ('a bit') > 'evolutive' ('a few/
little at a time') > 'iterative' ('several times')> 'incessative' ('incessantly')> 'celerative'
('quickly').
This reduplication is also used with relational nouns (denoting spatial relations) and adverbs. The meaning conveyed is attenuative with relational nouns
(57b), to be compared with (57a) and with the time adverb eutilem 'before' (58). The
meaning is iterative with the time adverbs epile 'afterwards' (59) and tsiketole 'when?'
(60):

(57)

a.
b.

moke=le
distant.location=Loc
'I will go far' (0:77)
mokeke=ma
distant.locationATT=LOC
'I will go a little way' (0:77)

le
FOC

ka-n-s

le
FOC

ka-n-s

go~FUT-lsg

go-FUT~lSG

Lule

623

(58)

euti[ti]lem
before[ATT]
'a little before' (G:83)

(60)

tsike[ke]tole=ma
when[ ITER]=INTG
'how often? (0:82)

(59)

epi[pi]le
afterwards[ ITER]
'from time to time' (G:84, V:85)

3 Conclusion
In this final section we will deal with two issues: the formal devices used in Lule evalua"
tive morphology and the question of the relationship between evaluative morphology and
word classes.
We have seen that Lule evaluative morphology uses two formal strategies: suffixation
and reduplication. As far as suffixes are concerned, we described one celerative suffix
(Section 2.1 ), one ameliorative suffix (Section 2.2), one pejorative suffix (Section 2.3),
one derogatory suffix (Section 2.4}, one augmentative suffix (Section 2.5), two intensive
suffixes (Sections 2.6 and 2.7) and one approximative suffix (Section 2.8). This is quite an
impressive array of evaluative suffixes, but as indicated in note 3, it is not impossible that
the ameliorative, the pejorative and the augmentative should be analysed as separate words
rather than as suffixes. If one chose to do so, the number of suffixes would be reduced to
one celerative suffix, one derogatory suffix, two intensive suffixes and one approximative
suffix, which is a less impressive system.
, As far as reduplication is concerned, we have described it as a polysemic process with
five evaluative meanings: attenuative, celerative, iterative, incessative and evolutive
(Section 2.9.).
Interestingly, some of the evaluative suffixes also have at least one non"evaluative
meaning. Added to a kinship term, the ameliorative, augmentative and one intensive suffix
(-ike) indicate that the referent ofbase noun is a blood relative (13), (21), (22) and (39). The
ameliorative may form with state and activity verbs derivates that express the meaning 'be
good to v' or 'worthy ofv-ing' (6)-(11). Finally, it is also noteworthy that the derogatory
suffix acts as a deverbal nominaliser in (18) and (19).
We now turn to the relationship between Lule evaluative morphology and the word
classes it attaches to. The celerative attaches to verbs. The ameliorative, the pejorative
and the derogatory attach to nouns and verbs. The augmentative attaches to nouns, state
verbs and one adverb. One intensive suffix (-ike) attaches to state verbs, one adverb, one
noun plus an element that could be a further noun or an activity verb. The second intensive
suffix (-ewi) attaches to state verbs. The approximative attaches to adverbs. The evaluative
reduplication is found with verbs of different aspectual classes plus nouns and adverbs.
Because of the small number of attestations of each sort of evaluative morpheme, it is hard
to draw definite conclusions about the relationship between evaluative morphology and
word classes. It does appear that evaluative morphology favours verbs and particularly
state verbs more than nouns. It is also more common with adverbs than with nouns.

Notes
1. A language that at present has only one or two surviving semi-speakers (Oolluscio and Gonzalez
2008).

624

Raoul Zamponi and Willem J. de Reuse

2. The remaining documentation of Lule consists of a short vocabulary by the Catalan Jesuit
missionary Jose Ferragut (Gilij 1782, 364-6)~ a text ofthe Lord's Prayer (Hervas y Panduro 1787,
I 02-3), not identical to that in Machoni ( 1732, Catecismo part, 1) and a few words in eighteenthcentury Jesuit reports.
3. In our philological analysis of Maccioni's grammar, the issue of word boundaries needs to be
addressed. Maccioni's written word spaces are one criterion by which we can decide whether he
considered a fonn to be one or more words. Unfortunately, this criterion is weakened by the fact
that Maccioni is not very consistent in his writing of word spaces. and in such cases it is difficult
to detennine whether a fonn consists of one or of several phonological words. This issue is of
course relevant in detennining which evaluative strategies are part of morphology, and which are
separate words and therefore not relevant to a description of evaluative morphology. As a rule of
thumb, we considered those evaluative elements which were almost always written together with
the preceding word or stem to be suffixes (that is morphology)~ and we considered those evaluative elements which were almost always written separate from the preceding word or stem to be
separate words (that is, relevant to the syntax rather than to the morphology). This rule of thumb
is not very satisfactory, as shown by the qualification 'almost always'.
The presumed affixal fonns with the suffix written separately are the originals in (5) aquem ecy,
(6) caip eci, ( 16) ecquys eytl and (21) pepe yqueps.
4. Examples from the grammar portion of Maccioni's Arte will be indicated by the abbreviation
G, in parentheses, after their translation, while those from the Spanish-Lule dictionary portion
or Vocabulario are indicated by the abbreviation D. The abbreviations G and D will then be followed by the number of the page in which the example occurs. Note that Dis paginated separately
from G.
5. Example ( 17) is to be compared to Italian donnaccia 'prostitute', from donna 'woman~.