Gajurru wurlun. Bumbujardi wurlun jamba. There's a fire burning. Let's go and have a look. There must be a big group of people. Two eaglehawks, one with sight and another blind, are there. Two sisters. They're dancing. They're stirring up clouds of dust. More generally, our results also imply that specific grammatical descriptions may have to be revisited on this point, and that theoretical or typological work for instance, on non-configurationality should not take simple generalizations about NP structure in Australian languages for granted.
The rest of this article is structured as follows. Section 2 discusses the make-up of the sample we use in this study.
Section 3 presents the set of parameters we use for determining constituency status, discussing the rationale behind each parameter. Section 3. Section 4 analyses the results, showing that especially the parameters of word order and locus of case marking provide clear evidence against the idea that Australian languages generally lack NP structures. Section 5 zooms in on discontinuous structures, examining the motivations for discontinuity where they are available, and arguing that the existence of discontinuous constructions is not invariably an argument against NP constituency.
Section 6 wraps up with some conclusions, including the argument that it makes more sense to typologize languages on the basis of where and when they allow NP construal for elements that belong together semantically, rather than on a yes-no answer to questions of constituency or dis continuity. This genetic diversity is represented in the sample: we included 65 Pama-Nyungan languages, covering most subgroups and geographical areas , as well as 35 non-Pama-Nyungan languages, representing almost all families and isolates.
An overview of the sample can be found in Appendix A , showing the genetic classification and the grammatical descriptions used for each language. The genetic classification of the non-Pama-Nyungan languages is based on Evans b , except for Enindhilyakwa, which has been re-classified as Gunwinyguan, following van Egmond As mentioned in the introduction, the aim of this article is to study NP constituency in its own right, independently from the more general theoretical question of non-configurationality. In other words, we want to know whether elements that semantically belong together can be construed as one syntactic unit, i.
In order to do this, we break down the concept into a number of concrete parameters that define constituency, which can be checked across the sample in a consistent way. Obviously, the sources do not allow us to check these criteria exhaustively for all languages, but there are a number of criteria for which we have good information across the entire sample. We distinguish between external and internal criteria for constituency. External criteria, discussed in Section 3.
Internal criteria, discussed in Section 3. Before we move on to the analysis, a terminological note is in order. External criteria for constituency focus on the interaction of a constituent with the structure of the clause: where case markers are located Section 3. In addition to telling us if NEs are treated as one unit in the clause, in some cases these criteria also provide a clear delimitation of one of the edges of the NE. The marking of case in an NE is a first criterion that may tell us something about its status as a syntactic unit.
In our sample, we find three basic options see also Blake : 78—91 : marking of one element in the NE, marking of all elements, or no marking at all. The first option is for case to be marked on only one element of the NE, i. The selection of one element for case marking implies that the NE is in fact one syntactic unit, which is marked for its role in the clause through one of its constituent parts.
In addition, if case is marked at either the left or the right edge of the NE, then this also serves to mark one of the boundaries of the NP. An example is Yandruwandha, where the ergative case suffix is attached at the right edge of the NE, as in 4 , thus showing that the nominal and its modifier can be analysed as a single NP, with the modifier forming its right edge.
Another option is for case to be marked on each element of the NE, i. In itself, this does not tell us anything about constituency, because there can be more than one reason for word marking.
One reason may be that the elements are separate NEs in apposition, which have the same case marker because they have the same function in the clause. This is how Blake ; see also : 89—90 analyses the structure in 3 above from Kalkatungu not in our sample, but see Footnote 15 , repeated below as 5 : the demonstrative, the adjective, and the noun are analysed as three elements in apposition, each of which is a dependent of the verb, and therefore receives its case marker directly from that verb.
Another possible motivation for word marking may be that the elements of an NE have the same case marker due to a process of agreement within a single NP. In such cases, there is usually other evidence for constituency, as in Yingkarta, illustrated in 6 below. In this language, word order is quite fixed, with modifiers preceding the nominal head, which constitutes independent evidence for constituency see also in Section 4.
Moreover, case may also be marked on only one element of the NP in this language, which further confirms that word marking in this structure really is agreement rather than apposition of separate NPs. Next to phrasal marking and word marking, the third option is that case is not marked in NEs all. This is often the case in head-marking languages most of the non-Pama-Nyungan languages in the sample , where the core argument relations are marked on the verb, and corresponding NEs remain unmarked especially for core arguments, but possibly also non-core arguments or adjuncts.
Obviously, these options are not mutually exclusive. It is common to find languages that allow both phrasal marking and word marking, as already mentioned for Yingkarta above. Relative frequencies and functions of the two alternatives are discussed in more detail in Section 4. More generally, the locus of case marking is also one of the criteria for which good information is available across the entire sample, and thus will serve as one of the central criteria in our analysis in Section 4. The best-known example is when a language has an element that obligatorily comes in the second position of the clause, following the first constituent.
Evidently, this criterion is more limited in the sample than the previous one, as only some languages have such slots, but there are some famous cases like Warlpiri, where the verbal auxiliary has a fixed position as the second element in the clause, following the first constituent e. This implies that all elements occurring in the first position before the auxiliary have to be analysed as one constituent. Accordingly, in example 8 , the nominal wawirri and the demonstrative yalumpu , both preceding the second position auxiliary, must be analysed as forming a syntactic unit.
Obviously, this criterion only allows us to determine the constituency status of NEs occurring in this slot, but not in other positions see 2 above , so it is slightly less conclusive than the previous criterion. Even so, the existence of slots defined in terms of constituency in a particular language does suggest quite strongly that construal as an NP is at least available in this language.
This is the external criterion that is least widely applicable in our sample: most of the grammars provide little or no information concerning prosody. Still, as prosody can be crucial in distinguishing several types of constructions cf. While such criteria are often part of the basic toolkit of initial fieldwork, they rarely find their way into grammars, 5 which means they are difficult to apply to our sample, and they have not been used in this study.
In addition to the external criteria, there are also two criteria that probe the internal structure of NEs to diagnose constituency: contiguity, discussed in Section 3. The relevant criterion here is whether the elements of an NE are contiguous, i. When they are, the elements are most likely one unit though this is not necessarily the case, as they could also be analysed as several single-item NPs in apposition, see also example 5 above, and Sections 4.
When they are not contiguous, however, as in the Garrwa structure in 9 , this has often been interpreted as evidence against NP constituency. The question is, however, whether this always follows when a language has discontinuous structures. We believe that the presence of discontinuous constructions in a language does not necessarily imply that contiguous constructions in the same language cannot be analysed as genuine NPs see further in Section 5 on this argument.
Therefore, we will investigate discontinuity separately in Section 5 below.
- Youd Be So Pretty If . . .: Teaching Our Daughters to Love Their Bodies--Even When We Dont Love Our Own.
- Emerging Trends in ICT Security.
- Courses of Action for Enhancing U.S. Air Force Irregular Warfare Capabilities: A Functional Solutions Analysis.
- Death Waits in the Dark: The Senoi Praaq, Malaysias Killer Elite (Contributions to the Study of World History).
Word order is the most important internal criterion for constituency in this study, because we have at least some information for almost all languages of the sample. If NEs have a fixed word order in a language, this is evidence for constituency, in the sense that the existence of a clear internal structure for an NE points towards unithood. This is the case, for example, in Umpithamu, as illustrated in the NP template in 10a and the structure in 10b.
Flexible word order, by contrast, has often been regarded as one of the main arguments against NP constituency in Australian languages. If we look at it in more detail, however, word order flexibility is not as straightforward a phenomenon as it might seem to be: it covers a range of different types of flexibility, and conclusions concerning constituency status for the NE differ accordingly.
Wambaya language, alphabet and pronunciation
As we will show in Section 4. An example is Umpila, as illustrated in 11 , where the order of the head nominal and the modifier is fixed, while the determiners personal pronoun, demonstrative, quantifier, or possessive pronoun can occur at either edge of the NE, but not in between the head nominal and the modifier.
Umpila Middle Paman; Hill This can be called flexibility, but it does not point towards the absence of internal structure, and therefore also the absence of constituency. On the contrary, it preserves the edges of the NE, and therefore shows that the NE is one unit. There are, of course, also languages that show genuine word order flexibility for NEs, i.
This is really the only type of language where flexibility provides evidence against constituency. Two other criteria that are sometimes mentioned in the literature are gender and number agreement. However, it is not clear what they can tell us about NP constituency, as they mark dependency relations rather than constituency, and are not even limited to the nominal domain.
- Catalog Record: A grammar of Wambaya, Northern Australia | HathiTrust Digital Library?
- A Grammar of Wambaya, Northern Territory (Australia) - Rachel Nordlinger - Google книги!
- Rachel Nordlinger - Google 學術搜尋引用文獻.
- Consistency of Translation Techniques in the Tabernacle Accounts of Exodus in the Old Greek (Septuagint and Cognate Studies Series 49);
- [PDF] Small languages and small language communities 57 - Semantic Scholar.
The only instance where this type of agreement could be interesting is when it is tied to case marking and changes location along with it — in which case it really is an instance of the criterion of locus of marking mentioned in Section 3. Table 1 provides an overview of the parameters we will use in our study. As already mentioned, we have to distinguish between those criteria for which we have good information across a large part of the sample locus of case marking, word order, and contiguity , and those criteria for which we only have information in some languages.
In this section, we discuss the results of our analysis for four of the five criteria introduced in the previous section, and we show that there is in fact little evidence against NP constituency across the sample. In Sections 4. This is followed by a discussion of prosody and occurrence in diagnostic slots in Sections 4. In Section 4. Tables listing the detailed results of our analysis for each criterion can be found in Appendix B Tables B-1 through B For a summary view of the results, the reader is referred to Table 2 in Section 4. The final criterion, which relates to discontinuity, is discussed separately in Section 5.