4. Supporting evidence
4.2 Typological evidence
4.2.2 Under-marking
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
35
demonstrate the distribution of specific elements within languages. Nevertheless, we follow Her (2015)’s study on languages in SMATTI (Sinitic, Miao-Yao, Austro-Asiatic, Tai-Kadai, Tibeto-Burman, and Indo-Aryan) and speculate that this is related to political landscape. In other words, category 1 and category 2 are actually two different systems in competition, and due to their different historical development, whether in terms of migration or economy, the gender plus plural marker system became more prominent in Europe and other western countries. On the other hand, in Asia, numeral classifier languages are stronger culturally and economically, therefore most of the languages also adopted numeral classifiers. Further details are provided in the GIS section, but as a reminder this is only preliminary speculation. It is not the main focus of our study and further research needs to be done to verify this hypothesis.
4.2.2 Under-marking
Starting from category 3 to 5, we enter the zone of under-marking, meaning that among countability marking and noun classification, the languages involved only represent one or neither of the two functions. Within category 3, coutability is represented once by the presence of grammatical plural markers, however noun classification is absent since neither numeral classifiers nor genders are attested. As an example in (26), we compare German in (26a) which has three genders and Afrikaans in (26b) which has no grammatical genders.
(26) Occurrence of plural markers but absence of numeral classifiers and genders Ellis et al (2012:18-19)
a. German
b. Afrikaans
‧
For German, gender agreement is observed on the determiners taking various forms as die, der and das for feminine, masculine and neuter. On the other hand for Afrikaans all nouns use the same determiner die since there is no grammatical gender differences. The same logic applies for grammatical plural. This phenomenon in languages of category 3 seems to be an exception to our hypothesis, but following our discussion within the theoretical sections: lack of countability marking and noun classification is possible. Nevertheless it is expected to be in low frequency since it would be on the extreme of the equilibrium continuum between economy and expressiveness. Our data also supports this prediction since only 8% (13/155) of the languages (Afrikaans, Finnish, Georgian, Central Kanuri, Brahui, Kenuzi-Dongola, Basque, Lezgi, Imbabura Highland Quichua, Evenki, Zuni, Chukchi, Sie) display this phenomenon.
It may be questioned that the ratio of category 3 is not small enough (8%) to be insignificant. Moreover, we it would be interesting to provide an explanation for the absence of noun classification within these languages. In other words, why would these specific languages avoid noun classification by numeral classifiers or genders?
It is not the focus of this paper therefore we do not elaborate too much this issue here, however as a possible direction for future studies, we speculate that the absence of numeral classifiers and genders representing noun classification in the languages of category 3 is due to the presence of other elements already fulfilling this function. A plausible example would be morphological case marking on nominals, which is a system in language used to express the syntactic and semantic relationships between clausal constituents, e.g. dative, genitive, locative, accusative among others. Its semantic function is tracking device, while its syntactic form is grammatical agreement, which seems similar to gender systems. According to this deduction, if genders and case marking fulfill a similar function of classification, we expect them to show a tendency of no co-occurrence, as it is happening with numeral classifiers and genders. It is necessary to point out that whether the overlap of semantic function between numeral classifier and genders vs case marking and genders is at the same degree needs detailed investigation, but at the actual stage we leave it for further studies to elaborate. To provide a preliminary evidence for our speculation, by doing a quick cross-matching between our database of 155 languages with the research result of Iggesen (2013)’s study on number of cases on nominals among 261 languages, we realize that the average quantity of cases is much higher in languages of category 3 compared to category 1 and 2. In other words, the tendency shows that if a language already has noun classification carried by either genders or numeral classifiers, it tends to avoid a double marking of the same function via case marking. On the other
‧
hand, languages with nominal case marking already have the noun classification function fulfilled, therefore they generally do not have gender or numeral classifier systems. Detailed numbers are displayed in Table 7.
Table 7. Distribution of nominal case marking among category 1, 2 and 3 Category 1 Quantity of cases Category 3 Quantity of cases
Spanish 0 Brahui 10 Europe and Africa in general has minimal (if not at all) case marking, as in Southeast Asia and the Pacific, while South-America and New Guinea show a diverse picture.
By linking this with Map 1 about the distribution of numeral classifiers and genders, we realize that the areas with either one of these elements are exactly the zone mentioned by Iggesen with minimal case marking, as an example Europe and Africa are the strongholds of genders while numeral classifiers have their hot spots in Southeast Asia. On the other hand, the regions at the intersection between these two forces precisely include South-America and New Guinea. Once more, this speculation is not the main purpose of this paper nevertheless we still mention it as a possible direction for future research.
For category 4, only the function of noun classification is realized via the presence of genders while coutability marking is not attested since neither numeral classifiers nor plural markers are present in the language. This tendency is expected to be low, which is supported by our data: within the 155 languages, only 1% (1/155) shows this type of distribution. The language in question is Piraha, the indigenous language of the isolated Piraha people of Amazonas, Brazil. This language is also the only surviving dialect of the Mura language, estimated to have a speaker population of near 400 people according to SIL. Regarding noun classification, it is attested to display four (Corbett, 2013; citing Everett, 1986: 280-281, 283-284) or three (Audring; 2008:102) genders. This contradiction does not influence our hypothesis as long as the gender
1 It may be argued that Vietnamese has plural markers and should not belong to category 2, but we follow Her (2013:12) by viewing them as quantifiers rather than grammatical plural markers.
‧
language known without number, numerals or a concept of counting, as demonstrated in (27a-c) the speaker may only express a relative quantity via ‘a few’ or ‘a big pile of’instead of the precise number.
(27) How to express quantity in Piraha without numerals (Everett, 2005:8) a. ti 7itii7isi hoi hii 7oogabagai
I fish larger pred. want
‘I want {a few/larger/several} fish.’
b. ti 7itii7isi baagiso 7oogabagai I fish many/group want
‘I want {a group of/many} fish.’
c. ti 7itii7isi 7ogii 7oogabagai I fish big want
‘I want {a big/big pile of/many} fish.’
Following this observation, it is then logical that grammatical plural marker does not exist within Piraha, since the plural requires number differentiations. As demonstrated in (28), within a phrase there is no number agreement therefore the hearer is facing the singular/plural possibilities for each noun. In this sentence, two nouns are mentioned by the speaker: ‘Piraha’ and ‘evil spirit’, since for each of them we have two possibilities of singular/plural interpretation, the hearer has in total two times two equal four possible meanings for this sentence.
(28) Lack of grammatical plural marker in Piraha (Everett, 2005:7) hiaitiihi hi kaoaibogi bai –aaga
Piraha people 3rd person evil spirit fear -be
‘The Pirahas are afraid of evil spirits.’ OR ‘A Piraha is afraid of an evil spirit.’ OR
‘The Pirahas are afraid of an evil spirit.’ OR ‘A Piraha is afraid of evil spirits.’
Nevertheless, it is attested that the mass/count distinction is still present within Piraha, as shown via (29). In (29a) the speaker may use 7aaibai ‘many’ to qualify 7aoo1
‘foreigner’ which is a count noun, however he cannot use 7apagi ‘much’ in the same situation, as shown in (29b). The reason is that 7apagi ‘much’ can only apply to mass nouns such as 7agaisi ‘manioc meal’, as shown in (29c). Finally in (29d) we may observe vice-versa, mass nouns such as 7agaisi ‘manioc meal’ may not be combined with 7aaibai ‘many’ which qualifies countable noun. It is necessary to point out that
‧
this distinction is more consistently analyzed as the distinction between things that can be individuated and things that cannot, thus independent of the notion of counting.
(29) Mass/count distinction expressed in Piraha (Everett, 2005:8) a. 7aooi 7aaibai 7ao7aaga 7oi kapio7io
foreigner many exist jungle other
‘There are many foreigners in another jungle.’
b. */? 7aooi 7apagi 7ao7aaga 7oi kapio7io foreigner much exist jungle other
? ‘There are much foreigners in another jungle.’
c. 7agaisi 7apagi 7ao7aaga 7oi kapio7io manioc meal much exist jungle other
‘There is a lot of manioc meal in another jungle.’
d. *7agaisi 7aaibai 7ao7aaga 7oi kapio7io manioc meal many exist jungle other
*’There is many manioc meal in another jungle.’
As a summary for category 4 and Piraha, even though grammatical plural markers are not present within the language, the mass/count distinction is still partially fulfilled via other markers, therefore it is still partially in accordance with our main hypothesis that the two functions of countability marking and noun classification are present simultaneously.
For category 5, it is the most extreme case of under-marking since numeral classifiers, genders and plural markers are all absent in the languages involved. According to our hypothesis it is possible but expected to occur in low frequency since the language would be the most economic but the less expressive on the continuum. This supposition is once more supported by our data: only 3% (5/155) of the languages display this type of distribution, they respectively are Usan, Yidiny, Kayardild, Chimariko and Imonda. Of course we still need to find a plausible explanation for their particular behavior: regarding noun classification, we follow the hypothesis mentioned previously in Table 7 that another element is fulfilling this function instead of numeral classifiers and genders. A similar pattern is found with Kayardild and Yidiny being attested to have 8-10 case marking, which is higher than the average of optimal marking categories in our data. Moreover, we point out another interesting phenomenon: some of the languages involved do not have numeral classifiers however they do possess similar but even more rare categories, such as noun or verb
‧
Australia, we may observe that a typical noun phrase is composed of a specific noun and an appropriate generic term: the noun classifier, e.g. we do not say ‘the wallaby is standing by the black pine’, while the preferred version is ‘the animal wallaby is standing by the tree black pine’ (Dixon, 1982:185). This type of noun classifiers are defined as “free morphemes standing in a noun phrase, next to the noun itself or within the boundaries of the noun phrase with other determiners of the noun and they are crucially found independently of the operation of quantification” (Grinevald, 2000:64). Their major difference with numeral classifiers is that noun classifiers do not require the presence of numerals.(30) Noun classifiers in Yidiny (Dixon, 1982:186)
minya ganguul jana-ng jugi-il gabuma-la animal-ABS wallaby-ABS stand-PRES tree-LOC black pine –LOC
‘The animal wallaby is standing by the tree black pine.’
Following our hypothesis, since noun classifiers are a type of noun categorization device (Aikhenvald, 2000:81), we may speculate that they fulfill the function of noun classification in noun phrases. The similar situation is observed in (31) with Imonda, a Border language of Papua New Guinea, but the system involved is slightly different:
verbal classifiers. They are found on the verb root rather than in the noun phrase structure, however they still classify the noun serving as of the nominal arguments of the verb (Grinevald, 2000:67). Since they carry the function of classifying nouns, we may speculate that they also represent the noun classification function we proposed in our hypothesis. As an example in (31), the classifier occurs on the verb but categorizes the noun into various types such as l for edible greens (10a), pui for breakable objects (31b), u for small animals (31c), i for water (31d), among others.
(31) Verbal classifiers in Imonda (Seiler, 1985:120-121) a. iahaf ka-m l -ai -h -u
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
41
d. po ka-m i -ai -h -u water one –GL CL –give –REC –IMP
‘Give me some water!’
On the other side about countability marking, similar to what we observed in Piraha, the languages involved here actually display quite restricted numeral systems, partially explaining the lack of plural marking on nouns. As an example with Usan in (32), its original numerals combine base-two and bodypart-counting systems and go only up to ten.
(32) Restricted numeral system in Usan (Reesink, 1987:61)
1 gari ‘one’
2 Ombur ‘two’
3 ombur gari ‘two one’
4 ger ombur ger ombur ‘a two a two’
5 uben ig ue ‘hand side not’
6 uben ig ue ig gari ‘hand side not side one’
10 uben ig ue ig ue ‘hand side not side not’
We make the same observation in Yidiny, which only has a numeral system going up to five displayed in (33a). The similar logic applies for Kayardild, an Australian language, which is not attested to have a numeral system going higher than 10, as demonstrated in (33b).
(33) Restricted numeral systems a. Yidiny (Chan, 2015)
1 = guman, 2 = jambul, 3 = dagul, 4 = yunggan.gunyjii or mugungabi, 5 = mala
b. Kayardild (Chan, 2015)
1 = warngiida, 2 = kiyarrngka, 3 = burldamurra, 4 = mirndinda, many = muthaa
As a summary, categories 3-5 include the cases of under-marking, where one or both functions of countability marking and noun classification are not represented. This situation is expected to occur in low frequency according to our hypothesis, which is indeed supported by our data. The lack of noun classification may be due to the fact that this function is actually fulfilled by other means in the languages, e.g. case marking, noun classifiers, verbal classifiers among others, but further research needs to be done on this domain. Regarding countability marking, we realize that its absence