Tutorial Week 3 – Chi-Pin’s answer guide
Part A
Step 1
Let us start with the three-word sentences. We are going to make two assumptions here:
(1) Assumption 1: Basque is a language of fixed word order.
(2) Assumption 2: Every sentence must contain at least a subject noun (N) and a verb (V).
1 Gizona joaten da
5 Txakurrak joaten dira
11 Txakurra joaten zan
Table 1
Note that Txakurra and Txakurrak are similar, and -k is possibly a suffix.
Now we have two word classes and a tentative result of our analysis of the two-word sentences:
(3) Tentative Result 1: Class A = {Gizona, Txakurra}
Class B = {joaten}
Class C = {da, dira, zan}
Step 2
Now let us put four-word sentences in Table 2:
2 Gizonakin zaldia ikusten du
3 Astoa atzo joaten zan
4 Gizonak atzo joaten ziran
6 Zaldiakin gizona ikusten du
7 Astoakin zaldiak ikusten zuen
9 Txakurrakin astoak ikusten ditu
Table 2
In sentences 2., and 9., we have Gizonakin and Txakurrakin, which are similar to the Class A words
Gizona and Txakurra in (3), and -kin may be a suffix which can be attached to Class A words as -k is.
We can then assume that the new words in Table 2 which can take -kin and -k and which appear in
same columns as Class A words in (3) do also belong to Class A. We can also observe that both joaten
1
and ikusten end in -en and appear in similar places, so they may belong to the same word class. Now
we have:
(4) Tentative Result 2: Class A = {Gizona, Txakurra, Zaldia, Astoa}
Class B = {joaten, ikusten}
Class C = {da, dira, zan, du, ziran, zuen, ditu}
We are not sure about the word class of atzo, because we have not seen any case in which the word
atzo ends in -kin or -k.
Also, since in Sentence 2., 6., 7., and 9., there are two Class A words in each sentences, it is very
possible that Class A words are nouns.
Step 3
Now let us put five-word sentences in Table 3:
8 Gizonakin txakurra atzo ikusten zuen
10 Zaldiakin gizonak atzo ikusten zituen
12 Gizonakin astoak atzo ikusten zituen
Table 3
We have
(5) Class A = {Gizona, Txakurra, Zaldia, Astoa}
Class B = {joaten, ikusten}
Class C = {da, dira, zan, du, ziran, zuen, ditu, zituen}
In Table 2, it seems like that atzo can appear in a Class A slot since we have
(6) A A B C
and
(7) A atzo B C
But we can not be sure because of limited information. Now, with more information, let us figure out
whether atzo belongs to Class A. First we can observe that Class C words do not take Class A endings
-k and -kin, and unlike Class A words, they always appear after Class B words. This is a strong hint that
Class C words are not Class A words, i.e. nouns. Hence, Class B words are very possibly verbs. Since
Class C words are not nouns, sentences in Table 1 are with intransitive verbs and sentences 2., 6., 7., 8.,
9., 10., 12. are with transitive verbs according to our presumption (2). In transitive sentences, we can
observe that the first Class A words are attached by -kin and the Class B words are ikusten. In addition,
the verbs of all those intransitive sentences in Table 1 are joaten. It leads to the presumption that -kin is
2
a case marker, joaten is an intransitive verb, and ikusten is a transitive verb. In Sentences 3., and 4.,
atzo comes with joaten. Since we have only one incorrect sentences, at least one of Sentences 3., and 4.
must be correct. Hence, atzo is not a Class A word, and sentences in Table 3 are not ditransitive
sentences. Now we have
(8) Tentative Result 3: Class A = {Gizona, Txakurra, Zaldia, Astoa}
Class Bi = {joaten}
Class Bii = {ikusten}
Class C = {da, dira, zan, du, ziran, zuen, ditu, zituen}
Class D = {atzo}
Step 4
Class C words seem to be similar. In fact, there is a parallelism between them:
(9) da, dira, du, ditu
(10) zan, ziran, zuen, zituen
It leads us to the presumption that -ir- / -it-, -a, and -u / -ue- are affixes. Now let us look at the
distribution of the infix -ir / it:
1 Gizona joaten d-a
2 Girona-kin zaldia ikusten d-u
3 Astoa atzo joaten z-a-n
4 Gizona-k atzo joaten z-ir-a-n
5 Txakurra-k joaten d-ir-a
6 Zaldia-kin gizona ikusten d-u
7 Astoa-kin zaldia-k ikusten z-ue-n
8 Gizona-kin txakurra atzo ikusten z-ue-n
9 Txakurra-kin astoa-k ikusten d-it-u
10 Zaldia-kin gizona-k atzo ikusten z-it-ue-n
11 Txakurra joaten z-a-n
12 Gizona-kin astoa-k atzo ikusten z-it-ue-n
Table 4
In Table 4, we can observe that, whenever there is a Class A word marked by -k, the Class C word in
that sentence is marked by the infix -ir- / -it-, and the only exception is Sentence 7. Sentence 7. is the
incorrect sentence, and its correct form is either
(11) Astoakin ikusten zuen.
3
or
(12) Astoakin zaldiak ikusten zituen.
In addition, whenever the verb is intransitive, the Class C word in that sentence is marked by -a, and
whenever the verb is transitive, the Class C word in that sentence is marked by -u / -ue.
The summary of our analysis up to this moment is
(13) Result: Class A = {Gizona, Txakurra, Zaldia, Astoa} with affixes -k, -kin
Class Bi = {joaten}
Class Bii = {ikusten}
Class C = {d-, z-n} with affixes -ir- / -it-, -a, -u / -ue-
Class D = {atzo}
Part B
Step 1
First, by observing the new sentences, there seem to be more affixes like -ke, -ua, -u, and -t. The -en
in ikusten and joaten in Part A also seems to be an affix. Let us put the new sentences in Table 5 with
all possible affixes separated from the roots:
13 Gizona-ke-kin astoa-k ikust-en d-it-u-e
14 Astoa-kin txakurra-k maluskat-en z-it-ue-n
15 Astoa-ke-kin zaldia-k maluskat-en z-it-u-t-e-n
16 Gizona-ke-kin zaldia-k ikust-ua d-ir-a
17 Astoa-ke-kin zaldia-k maluskat-ua z-ir-a-n
18 Txakurra-k ikust-ua d-ir-a
19 Zaldia-ke-kin txakurra atzo maluskat-ua z-a-n
20 Zaldia-ke-kin txakurra maluskat-en z-u-t-e-n
21 Astoa ikust-ua d-a
22 Gizona-kin txakurra ikust-ua d-a
23 Zaldia-ke-kin astoa ikust-en d-u-e
24 Txakurra-k atzo maluskat-ua z-ir-a-n
Table 5
We already know that Class A words are very possibly nouns and Class B words verbs. In Table 4,
Class C words stand with verb, are rich in morphology, and show agreements with both nouns and
verbs. This suggests that Class C words might be auxiliaries.
4
Step 2
Now let us observe the distribution of the affix -ke in Table 5. The morpheme -ke effects the forms
of Class C words, i.e. auxiliaries, in two ways. In Sentences 13., 15., 20., and 23., the first nouns of the
sentences are attached by -ke, the verbs are attached by -en, and the auxiliaries have an additional -e as
a suffix added to the transitive marker -u or a additional -t- as an infix sandwiched in the transitive
marker -ue. In Sentences 16., 17., and 19., the first nouns of the sentences are attached by -ke, the
verbs are attached by -ua, but the auxiliaries are marked by the intransitive markers -a, and there are no
additional -e or -t- affixes. It shows that the morpheme -ke systematically effects the behaviors of the
auxiliaries, but its effect is also conditioned by the presence of the morpheme -ua. Since auxiliaries in
sentences without -ke also show no additional -e or -t- affixes, we can presume that the morpheme -ke
will cause the auxiliarie to have an additional -e or -t- affix only if the verb in that sentence is not
marked by -ua.
We have found in Part A that the Class C words have a -a morpheme in intransitive sentences and a
-u / -ue morpheme in transitive sentences. However, in Sentences 16., 17., 19., and 22., even there are
two nominal arguments, which means that they are transitive sentences, the Class C words have a -a
morpheme instead of the expected -u / -ue morpheme. In addition, in these cases, the verbs are marked
by -ua instead of -en. This means that -ua might be a suffix which let a transitive verb behave like an
intransitive verb, i.e. a suffix which reduces the valencies of the verbs. There is more supporting
evidence: First, one of the nominal arguments in transitive sentences with verbs marked by -ua
becomes optional. Second, auxiliaries in transitive sentences with verbs marked by -ua do not have
affixes -e or -t- even if the first nominal argument in those sentences are marked by -ke, i.e. auxiliaries
show less agreements with nominal arguments.
The morpheme -ke and -k look similar, and they have complementary distribution, We do not have
-ke and -k happening in the same word, and -ke is always followed by -kin. Hence -ke and -k are
allomorphs. We know that Basque is an ergative language. Since the nouns in Sentence 1., 3., 4., 5., and
11., which must be subject of the intransitive verb joaten, can be marked by -k or not, -k must not be a
case marker. Since -kin always appear in one of the nominal arguments of sentences with transitive
verb roots, -kin is very possibly a case marker. Since Basque is an ergative language, -kin is very
possibly the ergative case marker, which marks the “doer” of a transitive sentence. We can put the
agreements between the ergative nominal arguments N1, the non-ergative nominal arguments N2, and
the auxiliaries in Table 6.
5
N2-ø N2-k
intransitive sentences
(Vi-en or Vt-ua)
N1-ø-kin + Vt-ua
d-a
z-a-n
d-ir-a
z-ir-a-n
intransitive sentences
(Vi-en or Vt-ua)
N1-ke-kin + Vt-ua
d-a
z-a-n
d-ir-a
z-ir-a-n
transitive sentences
(Vt-en)
N1-ø-kin
d-u
z-ue-n
d-it-u
z-it-ue-n
transitive sentences
(Vt-en)
N1-ke-kin
d-u-e
z-u-t-e-n
d-it-u-e
z-it-u-t-e-n
Table 6
Since -k / -ke are not case markers, they might be grammatical number markers. Follow this
presumption, -ir- / -it- shows the agreements between auxiliaries and the numbers of N2, and -e / -t-
shows the agreements between auxiliaries and the numbers of N1, but only in transitive sentences.
Step 3
Now we have enlarged lexica:
(14) Result: Class A = {Gizona, Txakurra, Zaldia, Astoa} with affixes -k / -ke, -kin
Class Bi = {joat} with affixes -en
Class Bii = {ikust, maluskat} with affixes -en, -ua
Class C = {d-, z-n} with affixes -ir- / -it-, -a, -u / -ue-, -e / -t-
Class D = {atzo}
We have discussed that Class A words are probably nouns, Class Bi intransitive verb, Class Bii
transitive verbs, and Class C auxiliaries. Class D words are probably adverbs, since they are optional in
sentence patters
(15) N ( D ) V AUX
(16) N N ( D ) V AUX
and since they stand besides the verbs.
Step 4
Let us summarize our analysis:
6
(17) Result: N = {Gizona, Txakurra, Zaldia, Astoa} with affixes -k / -ke, -kin
Vi = {joat} with affixes -en
Vt = {ikust, maluskat} with affixes -en, -ua
Aux = {d-, z-n} with affixes -ir- / -it-, -a, -u / -ue-, -e / -t-
Adv = {atzo}
For nouns, we have morphemes -k / -ke, which might be grammatical number markers, and -kin,
which is an ergative case marker. We have verb endings -en and -ua, and -ua reduces the valencies of
transitive verbs. For auxiliaries, we have affixes -ir- / -it-, which shows agreements with -k of non-
ergative nominal arguments, affix -a, which indicates the intransitivity, -u / -ue, which indicates the
transitivity, and -e / -t-, which shows agreements with -ke of ergative nominal arguments only in
transitive sentences.
The phrase structure rule is
(18) S → (N) N (Adv) V Aux
Based on our analysis, we can predict that sentences (19) and (20) would be grammatical
(19) Txakurrakekin gizonak ikusten ditue.
(20) Txakurrakekin gizonak ikustua dira.
since they follow the phrase structure rule (18) and the agreements which we have summarized. We can
also predict that sentences (21) and (22) would be ungrammatical
(21) Txakurrakekin ditue ikusten joaten.
(22) Txakurrakekin gizonak ikusten ditu.
since (21) does not follow the phrase structure rule (21), and (22) does not follow the agreements which
we have summarized.
7