Papuan linguistics

INDONESIAN: Buku ini berisikan data tentang berbagai spesies tumbuhan liar maupun yang dibudidayakan dan terkategori asli atau endemik yang bernilai guna dalam kehidupan sosial, budaya, ekonomi dan religius masyarakat suku Marori. Pengkajian dilakukan dengan melibatkan tokoh tokoh kunci atau tetua adat suku Marori yang masih memelihara pengetahuan tentang tumbuh-tumbuhan dan pemanfaatannya secara tradisional di dalam kehidupan sehari-hari. Oleh karenanya, di dalam buku ini banyak terdapat istilah-istilah dalam bahasa Marori mulai dari penamaan tumbuh-tumbuhan sampai pemanfaatannya sebagai wujud kepedulian kami atas kelestarian pengetahuan tradisional itu.
ENGLISH: This book contains data about various species of wild plants as well as those which are cultivated and categorized as native/endemic, which are of high social, cultural, economic and religious values in the lives of the Marori people. The documentation research was undertaken with information collected from key traditional elders and leaders of the Marori tribe who still maintain deep knowledge of plants and their traditional use in daily life. This book contains many terms in the local Marori language, and rich information of different kinds reflecting the local concerns for the sustainability of traditional knowledge.

In the Northeast Halmaheran languages (Tobelo, Galela, Sahu, Modole, Pagu, and Tabaru) all nouns are immediately preceded by an article o or ma. These articles co-occur freely with other nominal morphology, with the exceptions of first and second person possessive pronouns. Since they occur on all nouns, both articles can function as ligatures within the noun phrases, particularly within possessive constructions and possessive-like attributive constructions. On simple nouns the function of the articles is less clear. This paper draws on a quantitative analysis of Tobelo narrative texts to argue that the articles function as markers of dependency relations in discourse. The article o is the default article used new or unimportant referents, while the article ma signals a dependency relationship between the following noun and another discourse entity, which may or may not be explicitly mentioned. This notion of discourse dependency not only confirms the intuition of early twentieth century observers but also helps to explain apparent head shift phenomena in complex noun phrases. This paper concludes with a discussion of relationship between Northeast ma and Proto-Malayo-Polynesian *ma, suggesting a possible additional path of contact between North Halmaheran languages and neighboring Austronesian languages.

Winner of the 2013 ALT Gabelentz Award --- Mian is a non-Austronesian (‘Papuan’) language of the Ok family spoken in the Highlands fringe in western Papua New Guinea. Mian has approximately 1,400 speakers and is highly endangered. This grammar is the first comprehensive description of the language. It is based on primary field data consisting of a text corpus that covers different genres of the oral tradition, namely myths and ancestor stories, historical accounts, accounts of the initiation ritual, conversations, and procedural texts. The corpus was recorded by the author during a total of eleven months of field work from 2004 to 2008. The book provides a thorough description of all areas of Mian grammar and gives an in-depth analysis of many points of typological interest, such as the complex system of lexical tone, the interaction between a gender system and a system of classificatory prefixes on verbs of object movement, manipulation or handling, which allows the highlighting of certain characteristics of a referent in a given situation, the complex verbal morphology which allows fine-grained tense-aspect-mood distinctions, and a switch-reference system in which switch-reference suffixes on medial verbs are homophonous with and derived from suffixes functioning as tense and aspect markers in final verbs. The book is rounded off by a collection of traditional and contemporary texts (fully glossed and translated) and a word list comprising some 1,600 items, giving lexical tone, word class and meaning.

The Torricelli language of Onnele (ISO 639: onr) includes a remarkable set of strategies to indicate number, and to express other related semantic notions of extendedness and pluractionality that go well beyond the paradigmatic categorization of number. Nominal plurals are rare outside of kinship terms, and the number of any particular referent is generally inferred from context. For object referents, the normative strategy is little interested in precision of count, which stands in stark contrast to the geographically closest Torricelli language of Olo, which includes over fifty subclasses of nouns based on number markings (Staley 15). The Onnele counting system only includes ‘one’ (wokera) and ‘two’ (plainre), and then combines these (e.g. plainrikera ‘three’). Nevertheless, Onnele exhibits over fifteen different strategies to indicate some kind of number.
Number is shown in a wide variety of ways among Onnele nouns, adjectives, possessive pronouns, question words, verbs, conjunctions, and pronouns. A compound noun includes each of the singular parts to form the resulting plural complex (e.g. ino ‘younger.brother’ + rani ‘older.brother’ = inrani ‘brothers’). This strategy allows for greater disambiguation than a mere singular/plural distinction, for another construction adds a plural suffix to some kinship terms (e.g. rani ‘older.brother’ + -mpi ‘PL’ = ranimpi ‘older.brothers’). Many kinship terms use different roots for singular and plural (e.g. awa / mokapi ‘mother’s.younger.sister(s)’). A few nouns can show plurality through partial or full reduplication, yet this often shows plurality in the sense of a large degree, or in a distributive sense. The same is true of adjectival suffixes. Possessive pronouns can sometimes indicate the number of the possessed noun, but this is still mostly determined from context. Question words can indicate number (e.g. fina ‘who.SG’, fopi ‘who.PL’). Most verbs are regularly marked with a portmanteau subject agreement morpheme across a paradigm that distinguishes singular and non-singular. Some verbs can also be marked for object number agreement, and these suffixes exhibit the greatest divergence of number-marking concord among the closely related Onnele dialects. The Onnele pronoun system exhibits a complex system of number – almost an open class in its use of the ‘and’-verb to form compound pronouns – in order to track numbered referents among dual forms as well as any other combination of numbered referents (e.g. wuyape ‘3SG.and.2PL’). A verbal prefix yem- conveys some kind of pluractional meaning of reciprocity or coordinated action among its actors, depending on the lexical and contextual constraints. Finally, separate verb lexemes are sometimes used to indicate whether the action is done to one object or to a plurality of them.
The semantic category of number is much more complex than simply identifying the singularity or plurality of nouns. Rather, number may include various semantically related notions to plurality across many word classes, and this complexity deserves more attention, as indicated by such recent typological volumes as Storch and Dimmendaal 2014 and Mattiola 2019. For the Onnele data, I need to look further into those nouns that do exhibit nominal plural distinctions and classify them accordingly. Also, some number markings were previously labelled as ‘plural’ when some other kind of extendedness or pluractionality may really be at play (e.g. adjective suffixes). The distinction between ompla ‘thing’ and ommo ‘things’ may really be something other than singular/plural since local speakers frequently use them interchangeably for singular and plural referents. Since this noun can also be used in various senses to indicate ‘food’ and ‘work’, perhaps the distinction is more along the lines of countable vs. uncountable.

Across traditional Papuan languages and cultures, personal names are special. The principles of name-giving are in many ways as diverse as the languages and the cultures themselves. Naming systems display 'both linguistic intricacy and cultural complexity' (Kroskrity 2021). This contribution addresses the types of names, including name-tunes, and their linguistic properties.

This is a grammatical description of the Papuan language Iatmul. It is based on almost fourteen months of fieldwork in Papua New Guinea, distributed over four trips in 2005/2006, 2007, and 2008. The structure of the book is as follows: Chapter 1 presents the language, its genetic affiliation, and its typological profile; the area where it is spoken and its sociolinguistic situation; and the fieldwork setting. Chapter 2 describes the phonology. Chapters 3-8 deal mostly with morphological issues, while chapters 9-15 describe phenomena pertaining to both morphology and syntax. Chapters 16-18 are about interclausal relations. Chapter 19 discusses verb-argument relations and the order of constituents in the clause. Chapter 20 goes beyond morphosyntax and discusses two semantic areas of cross-linguistic relevance, the versatility of speech report constructions, and the extreme polysemy of certain lexical roots. Chapter 21 deals with sociolinguistic topics, strategies for lexical enrichment on the one hand, and language shift and its structural correlate, language obsolescence, on the other. The appendix includes six fully glossed, translated, and annotated texts, a short word-list, and a list of all affixes.

While linguistic terminology is a meta-language used to describe object languages, it can itself be an object of study. It then becomes obvious that quite often scientific terms have semantic properties similar to those of the everyday object-language expressions they are supposed to describe. There is often a narrow and a broad definition, or a core meaning with one or more extensions. I will present a few common terms from linguistics to illustrate that linguists have not reached agreement on basic concepts.

The ‘subordinate clause’ is such a concept. For some it is defined as ‘embedding’, whereas others use it as a synonym of ‘dependent clause’; but the latter two concepts are not used univocally either. Sometimes the subordinate clause is defined as a subtype of dependent clauses, sometimes it is a subtype of embedded clauses, and sometimes it is exactly the other way round. And while some linguists stress morphological and/or syntactic criteria when describing a clause as ‘subordinate’, others focus on its semantic and discourse properties.

This conceptual confusion makes it difficult to describe the Iatmul suffix -ka, the core function of which is to mark a verb form as non-finite. It appears in a variety of contexts ranging from coordinate-like clause adjunction to simple adverbs and periphrastic inflection. These can be located on a CONTINUUM OF SYNTACTIC INTEGRATION, of which subordination and embedding occupy overlapping stretches.

This is a grammar of Ma Manda, a language of Papua New Guinea, which covers major aspects of this previously undescribed language. The analysis is supported by culturally-embedded examples from a recorded text corpus. The result is a comprehensive preservation of this endangered language for its speakers, and for linguistic and anthropological scholars working in the Papuan arena.

The languages of Timor, Alor and Pantar (TAP) are notable for their object agreement prefixes. Previously, this has been highlighted because this exists largely without subject agreement (a rare pattern crosslinguistically; Klamer 2014, Siewierska 2011) and the proliferation of different prefix series and the semantics they express (Fedden et al 2014, 2013, Kratochvíl 2011, inter alia). One particular feature has not raised much comment, despite its rarity and difficulties it raises for syntactic theory (Chumakina & Bond 2016): object agreement is only obligatory for a lexical class of transitive verbs.
This is particularly unfortunate since classes defined in the same way are a feature of many Trans-New-Guinea (TNG) languages, the prefixing class labelled object verbs, even being reconstructed to the protolanguage (Suter 2012). They exist in a number of non-contiguous groups of families: Dani; Ok and Anim; Kainantu-Goroka and Finisterre-Huon. These languages have dealt with this uninflectability in different ways, through support verbs, that resemble auxiliaries, or free object pronouns. What they all share are cognate agreement prefixes based on the TNG pronominals (see Suter 2012 cf: Ross 2005). The TAP languages also look to have object verbs defined by prefixes that may well be derived from these same pronominals.
This connection between the TNG and TAP is especially significant since the pattern is crosslinguistically rare. This implies that it is unlikely to have been caused by chance. This provides important extra evidence of TNG-TAP interconnectedness. Moreover, it is a serious question whether these lexical verb classes would likely be diffused and apply to the whole transitive verbal lexicon of TAP (conjugation classes are not likely to be recombinantly borrowed; Panov 2015, Koutsoukos 2016, or even Robbeets 2015, 2017). This may leave inheritance as the most probable explanation for why object verbs are found in both TAP and TNG. This would then add to the growing evidence for the TNG descent of TAP.

This is the seventh addendum to my manifesto "Rapid Progress in the Genetic Classification of the World’s Languages ". It contains a list of almost 200 useful books I've collected about the world's Language families. It is obviously not exhaustive, and it includes books only (not papers or online resources).

The classification of languages as Austronesian is sometimes carried using a rather generous view of cognacy. A few well-known PMP cognates and a perception that an island was uninhabited prior to the Oceanic expansion have often allowed linguists to assign languages to Austronesian without serious historical argument. Blust has drawn attention to Kokota and the Whiteman languages, but this is clearly also a problem for the languages of Vanikoro and Utupua, small islands in the Southern Solomons. These languages show very limited cognacy with proto-Oceanic, and often with irregular correspondences. If they are Oceanic, they have certainly developed in ways quite at odds with almost everything we know about diversification in Austronesian. However, there is no archaeological evidence for prior settlement, so this cannot be a situation of language mixture such as may be the case in the Solomons. A possible explanation for this is these languages result from intensive interaction with a variety of NAN languages in the Oceanic homeland, and that the early canoes carrying settlers had crews already representing an ethnic and linguistic mosaic, which was transferred to the new settlements. The paper also considers if a similar model makes sense for other problematic languages in the region.

THIS IS AN EARLY DRAFT OF THE PAPER AND NOT THE FINAL VERSION. PLEASE GO TO http://www.revue-archipel.fr/revues/archipel90.html TO GET THE FINAL VERSION OF THE PAPER.

This paper defines a Wallacean linguistic area occupying almost the same geographical area as the zoological area of Wallacea. Just as zoological Wallacea is a complex zone of transition between two distinct faunal types, Asian and Australian, so Linguistic Wallacea straddles two quite different linguistic regions, Austronesian and Papuan. Despite their diversity, however, both the Austronesian and the Papuan languages of Wallacea display a shared set of features which distinguishes them both from the Austronesian languages to their west, and from the Austronesian and Papuan languages to their east. The nature and dispersal of the features shared by Wallacean languages indicates that in prehistoric times, prior to the Austronesian language dispersal, a large part of eastern Indonesia was already inhabited by seafaring, agricultural peoples speaking Papuan languages.

This booklet contains a Iatmul–English dictionary based on my fieldwork research in Korogo (East Sepik Province, Papua New Guinea) in May 2005 and from October 2005 to June 2006.
The dictionary entries are presented in the alphabet that I designed to write the Iatmul language. Although the writing system is based on previous attempts to write the language, I had to make a few modifications. Most of these were discussed with and approved by Iatmul speakers. An overview of the sounds and the spelling is given in the section ‘The Iatmul Writing System’. The first part of the dictionary is arranged according to the order of the letters in the English alphabet. In the second part, nouns from the same semantic domain are grouped together. The first part helps to look up Iatmul words. The second part helps learners to get acquainted with words that are related to each other by
meaning.

Papua Barat adalah salah satu Wilayah yang Belum Berpemerintahan Sendiri (Non Self Governing Territories) berdasarkan Resolusi Majelis Umum PBB No. 448 tanggal 12 Desember 1950 yang berada dibawah Administering Power Kerajaan Belanda sejak 1945 - 1962 kemudian ditransfer administrasinya ke Indonesia melalui suatu Referendum ilegal yang disebut PEPERA tahun 1969 kemudian Indonesia tidak melaporkan proses Self Government setiap tahun di Sidang Umum PBB sesuai Piagam PBB Pasal 73e

A verbal inflection whose primary function is to indicate future time reference is commonly called ‘future tense’. In descriptive practice, however, such morphological markers are often polysemous or multifunctional. In the main part of this paper, I will present three descriptive case studies illustrating some of the problems involved. The choice of these three case studies – Turkish, Basque, and the Papuan language Iatmul – is motivated by the fact that they have markers that are often glossed as ‘future’ – including by myself in the past – but a closer examination will reveal that this label is problematic in all three languages. It is therefore avoided in this paper, the more appropriate labels ‘prospective’ (for Turkish and Basque) and ‘irrealis’ (for Iatmul) being advocated instead. In order to facilitate the understanding of the data, the language-specific sections 2 (Turkish), 3 (Basque), and 4 (Iatmul) will begin with a short overview of the verbal morphology of each language, limited to the aspects relevant for the subject matter of this paper. In the second part of sections 2-4, the approach is first onomasiological, that is, we start with the function and look for the corresponding structural means to express that function in a given language. In our case, this means that we single out the concept of future time reference and look for its grammatical correlates in our language sample. The structural means found through this approach will be called FUTURITY MARKERS, intended as a mere “working label” for a COMPARATIVE CONCEPT in the sense of Haspelmath (2010: 664), which can be used for cross-linguistic comparison without making any claims about the exact morphosyntactic or semantic properties of the different markers. Given that these properties are different depending on the marker and the language, the second step is semasiological (form-to-function), as we describe the polysemy and, where applicable, diachrony of the morphemes and constructions found through the onomasiological approach. Section 5 draws attention to the fact that dependent clauses can have TAM distinctions and markers different from those of main clauses. This will be illustrated with nominalized predicates in Turkish and Iatmul. The corresponding analytical and terminological issues will also be addressed. Section 6 deals with overlap and differences between prospective, future, and irrealis. We will see that the distinction between prospective aspect and future tense also depends on the extent to which a prospective construction is grammaticalized. While a present prospective can develop into a future, past prospectives tend to take over the function of counterfactual conditionals in the past. Both scenarios involve a shift from realis to irrealis status. This will be illustrated by strikingly parallel phenomena in Turkish and Basque. Section 7 will sum up the discussion. First of all, we should observe a distinction between futurity markers and future tense markers. If we follow the argumentation in Haspelmath (2010: 671), a given language would then have futurity marking as a (cross-linguistic) comparative concept, but it would not have future tense as a (language-specific) descriptive category. This is comparable to those languages that use verbs (as defined by language-internal criteria) “as” adjectives (in the sense of a comparative concept). We will see in this paper that Turkish and Basque use their aspect categories, and that Iatmul (and English) use language-specific modal markers to express a cross-linguistic category future (tense/time). However, aspectual, temporal, and modal markers are connected by grammaticalization scales, so that in addition to language-internal structural criteria, we also have to take into account that tense markers may develop from aspect markers. All TAM marking contributes to the SITUATION PERSPECTIVE, where zooming in corresponds to the individuation of the predicate, analogous to the referential individuation of participants.

This ethnographic study of folk biology among the Tobelo (a West Papuan-speaking ethnic group of Halmahera Island, Maluku, Indonesia) outlines local cultural presumptions about classifying flora and fauna, describes the system of folk biological nomenclature in terms consistent with the morphology and syntax of the Tobelo language, and analyzes the local system of folk classification within a posited semantic domain of "biotic forms."
    In the local linguistic context, dialectal differences, multilingualism, an apparently strict in-law name taboo, and particular speech registers for which Tobelo consider their own language inappropriate are shown to affect word formation, the adoption of foreign plant and animal names, and other aspects of ethnobiological classification. Culturally, the belief that names for plants and animals were set down by ancestors vastly more familiar with local biota than are their descendants, the notion that there is a "proper" name for virtually all easily visible plants and animals, and that much knowledge is and should remain esoteric, justify several alternative ways in which the Tobelo may reconcile individual or dialectal variation to determine "proper" details of classification consistent with these presumptions.
    Nomenclature is considered in detail. The importance of recognizing the lexemic status of homonymous and polysemous terms is illustrated; and means of recognizing lexemes having the same form as non-lexemic expressions are detailed. A morphosyntactic classification of lexemic types is here applied to the formation of terms in this domain.
    Unlabeled classes within this semantic domain (including the highest-level class BIOTIC FORM) are posited, and new methods are presented for determining and evaluating such "covert categories." A critique of other procedures based on perceived similarities among plants and animals shows that the only local cultural significance of those classes may be their sudden appearance as a result of tests designed to find them, that similarities observed may not be those used in hierarchically relating folk taxa, and that such classes do not in any case belong in a linguistic description. From a systematic review of Tobelo lexemes it is possible to avoid these difficulties.
    The analysis of Tobelo folk biological classification (the system of semantic relations among usually lexically labeled classes) provides various types of evidence for the distinctiveness of a "basic" level, and details methods for distinguishing basic terms. Taxonomic relations order the set of hierarchically related folk classes into eleven levels: the widest or "basic" level, along with six above and four below. on-"regular" elements of this folk taxonomy include nonsymmetric and disjunctive contrast, "residue" of higher-level classes, ambiguous subclass-superclass relations, and dual structural positions of a single class in the overall hierarchic structure.
    Also analyzed are other types of semantic relations among folk classes, including a 'mother'-'child' relation among FAUN AL FORMS, crosscutting and intersecting subclasses of the basic class, and classification by growth stage and size. The Tobelo are able to use these methods of classifying local fauna and flora to justify the observed differences among themselves (or among Tobelo dialects) in how they classify the same plants and animals. They even sometimes use this classificatory system to productively predict the existence of plants and
animals that have not yet been observed, much as our chemists once used the Periodic Table of the Elements to predict the existence of elements that had not been observed.
    Detailed folk classificatory, nomenclatural, and systematic botanical and zoological information for all recorded BIOTIC FORMS is given in the appendixes, which are based upon extensive collections of Halmaheran terrestrial and marine animals and plants, along with their associated ethnographic information.

The Huon Peninsula languages are a family of Papuan languages spoken in the northeast of the island of New Guinea. The family comprises 21 languages. One of them, Kovai, is not spoken on the Huon Peninsula but on Umboi Island in the Vitiaz Strait, which separates New Guinea from New Britain. In this dissertation, the Huon Peninsula languages are compared to each other with the aim of reconstructing the commonly inherited morphology. The verb is rich in morphology, showing prefixes as well as suffixes. The prefixes occur on a small closed class of verbs and index the person and number of the object referent. About two dozen verbs with object prefixes can be reconstructed to various subfamilies and five of them to the top-level family. The verb suffixes index the person and number of the subject referent and express tenses or moods. Partial paradigms for five tenses and moods can be reconstructed to Proto-Huon Peninsula and several more to the subfamilies. As in many other language families, the personal pronouns are stable so that all forms can be reconstructed. Other pronouns that lend themselves to reconstruction are the possessive affixes and basic as well as elevational demonstratives. In the nominal domain, the Huon Peninsula languages share six types of case enclitic, three of which can be reconstructed to the top-level family. Apart from reconstructing aspects of the proto-languages of the languages compared, the dissertation also sheds light on processes of language contact that have shaped them.

Onnele is the local language name for Northern One [onr], an undocumented, endangered Papuan language in the West Wapei subgroup of the Torricelli family in Sandaun Province, Papua New Guinea. The Torricelli family includes around fifty languages, yet accessible descriptions are lacking, especially for the West Wapei subgroup.
Using a corpus of natural texts reflecting various discourse types, I seek to narrow the typological Torricelli research gap with a data-driven grammar description of Onnele, particularly the Goiniri dialect. While I may occasionally highlight some remarkable contrasts between the related Onnele dialects of Wolwale and Romei-Barera, this thesis will focus on the Goiniri dialect.
This research will include description of the Onnele context (geographical, historical, cultural, and sociolinguistic setting), phonology (phonemes, syllable structure, morphophonemics, and prosody, including tone), morphosyntax (word classes, noun and verb phrases, adpositional phrases, clauses and sentences), and functional grammatical categories (number, possession, tense, aspect, mood, accompaniment, directionality, pluractionality). The thesis culminates with a description of text relationships beyond the sentence at the level of paragraph and discourse. Interesting Onnele features, with atypical Torricelli features marked [!] include:
• min. phoneme inventory – no voicing contrasts [!]
• lexical tone contrasts [!]
• min. residue of noun classes on ADJ & kinship [!]
• 12+ strategies for marking plurality
• clusivity distinction in pronouns [!]
• ‘and’-verbs (cf. Brown & Dryer 2008)
• remarkable complex pronouns
• non-configurational noun phrases (Donohue 2000)
• prepositions and postpositional locative nouns
• 5 possessive constructions?
• no grammatical tense
• 10+ verbal aspects
• regular verb serialization without clause chaining
• topicalization particle
• demonstratives also pointing to time and text
• surface structure marking at discourse peaks

The impersonal experiencer construction is well known among the Papuan languages of New Guinea, regularly appearing in typological descriptions (Foley 1986; Foley 2000). This has two regular features: an idiomatic nominal as a possible subject and an experiencer formally coded as an object. In one particular Trans New Guinea language, Yagaria of Eastern Highlands, this is true only on the surface; other parts of the grammar reveal this experiencer to be a subject like any other. There is some evidence this mismatch of behaviour and form derives from an earlier stage of the language where the experiencers were unequivocal objects. This development from object to non-nominative subject meets the suggestion of Cole et al. (1980). There is, however, a significant difference from the instances previously described. This is the presence of noun-verb complex predicates. This has allowed the idiomatic nominal to be reanalysed as part of the predicate and lose its subjecthood. Thus this was not just a change in grammatical relations: the erstwhile transitive clause was reanalysed as intransitive and the experiencer, the only argument, as subject. Both these noun-verb complexes and the impersonals themselves are modelled in the formal syntactic framework of Lexical Functional Grammar (Bresnan 2001).

Fedden, Sebastian. 2020. Grammaticalization in Mountain Ok (Papua New Guinea). In Walter Bisang and Andrej Malchukov (eds.), Grammaticalization scenarios. Areal patterns and cross-linguistic variation. A comparative handbook. Comparative Handbooks of Linguistics 4, 1007-1041. Berlin: De Gruyter Mouton.

This paper discusses the morphosyntax of valency in Nen, a language of the Yam family of Southern New Guinea, with particular reference to the interactions between case morphology and verbal agreement (up to two arguments indexed). It is based on the set of predicates investigated in the Leipzig Valency Project.

This book contains the stories that the Iatmul people told me during my fieldwork research on their language in Koloku (or Korogo, Korogu; East Sepik Province, Papua New Guinea) from October 2005 to June 2006, and from July to September 2007.

In this paper, we seek to draw attention to Malayo-Polynesian languages outside of the Oceanic subgroup with innovative bases and complex numerals involving various additive, subtractive, and multiplicative procedures. We highlight the fact that the number of languages showing such innovations is
more than previously recognized in the literature. Finally, we observe that the concentration of complex numeral innovations in the region of eastern Indonesia suggests Papuan influence, either through contact or substrate. However, we also note that sociocultural factors, in the form of numeral taboos and conventionalized counting practices, may have played a role in
driving innovations in numerals.

The Finisterre-Huon stock was established by Kenneth . His evidence for a genealogical relationship among these 60 plus languages is reviewed and found to be compelling. In this article I take up one of his pieces of evidence, verbs with pronominal object prefixes, and elaborate on it.

The author uses a 'canonical' approach to offer a new perspective on the complex phenomenon of inflectional classes. This means extrapolating from what there is to what there might be, in order to define the theoretical space into which real instances fit. To do this, the author proposes eight criteria, grouped under two overarching principles. These are: I. distinctiveness: canonical inflectional classes are as clearly distinct as possible; and II. independence: the distribution of lexical items over canonical inflectional classes is unmotivated. The author investigates the various deviations from these principles, by considering in turn the more detailed criteria which exemplify them. While one might reasonably expect that 'canonical inflectional class' is an ideal without exemplars, the author finds an example which comes remarkably close to canonical.

The plosive inventory includes both voiceless and voiced plosives at the bilabial, alveolar, and velar places of articulation. (See below for discussion of the glottal stop.) The voiceless stops lack any significant aspiration. Voice onset time is close to zero for word-initial voiceless stops. In intervocalic position more aspiration may be present, though voice onset time is typically less than 15% of closure duration. Voiced stops are fully voiced throughout. Place of articulation and voicing contrasts in stops are indicated in the following examples.