Wolfgang B. Sperlich, ed. 1997. Tohi Vagahau Niue/Niue Language Dictionary: Niuean–English, with English–Niuean Finderlist. PALI Language Texts: Polynesia. Honolulu: Government of Niue and University of Hawai‘i Press. 596 pp., $45.00 cloth.

The Polynesian language of Niue has up till now been poorly served for descriptions and lexicography, in spite of its cultural and historical importance, so this volume is a very welcome addition to the gradually increasing inventory of substantial bilingual dictionaries of Polynesian languages. Niuean is with Tongan one of the two representatives of the Tongic branch of the Polynesian language family. There is no comprehensive grammatical description, and McEwan’s 1970 Niue dictionary has long been in need of expansion and revision. Approximately 12,000 Niueans live in New Zealand—about 85% of the total number of Niuean speakers. Given the problems of language maintenance that inevitably affect immigrant populations of 2–3 generations, this dictionary will be an invaluable educational resource for that group, as well as for the population remaining in the homeland.

The Niue language dictionary (NLD) was the work of a committee of seven Niuean speakers chaired by Atapana Siakimotu, and an editor, Wolfgang Sperlich. Bruce Biggs was a consultant to the project. I shall refer to this group of people as "the editors." From its striking blue and gold dustjacket incorporating Sale Jessop’s handsome illustration, to the clarity of the page layout and typefaces, this is a production of which the Government of Niue and the Dictionary Project can be proud. It should find a place on the shelves of every Niuean household and of anyone who is interested in Polynesian languages. With 10,000 word entries, and a finderlist containing 4,000 English headwords, it becomes one of the most substantial Polynesian language dictionaries to date. Although it does not approach the size of the Hawaiian dictionary with its 26,000 word entries (Pukui & Elbert 1986), it surpasses it in the comprehensiveness of its entries, in particular the abundance of Niuean sentence and phrase examples. These have been composed by the committee rather than derived from texts, in order to show clearly the uses of the words they illustrate (sometimes to the extent of making it apparent that the word-class information given in the entry is incorrect, a point to which I will return below).

A perennial dilemma for bilingual dictionaries is the question of intended readership. The editors of NLD confront this matter early with the statement "The dictionary was written by Niueans for speakers of Niuean as a repository of their language" (1), and go on to say that the organization of entries was devised for the convenience of native speakers, not primarily for learners of Niuean and comparative linguists. The consistency with which this aim is pursued, then, will be one measure of the success of NLD. [end page 193]

The main consequence of this policy was the formatting decision to include all derivations of a Niuean word as subentries within the entry for the headword, dispensing with strict alphabetization and with cross-references. Whereas McEwen (1970) followed this format for highly productive derivations such as reduplication or prefixation with faka-, and for other derivations which do not involve an infringement on alphabetization, NLD uses it also for idiosyncratic compounds and for less productive forms of prefixation, such as ma- prefixation. So, to take an example at random, let us consider the expression oneone feutaaki ‘quicksand’. In McEwen’s Niue dictionary it appears as a subentry under the headwords feutaaki, uta¹, and one. In NLD the phrase appears only under one (and the entry for the derived lexeme feutaaki appears only as a subentry under uta¹). The economy of this decision is obvious and on the whole it should work well for readers who understand the morphological composition of Niuean words. Moreover the system is explained with admirable clarity in the user notes. The one major disadvantage that I noticed concerned abbreviations. The word naga ‘relative, relation’, an abbreviation of matakainaga, is buried deep within the entry for mata¹, as is matakainaga. (In McEwen, naga is a headword with the gloss ‘bosom friend’). A nonspeaker of Niuean, no matter how astute, would have no way of locating naga in this dictionary. In this case a separate entry and cross-reference should have been used.

Another consequence of the decision to put all derived words under the headword is the often unwieldy size of entries, and one suspects that this might have been the reason to list mata¹ n. ‘face, eye’, as an entry separate from mata² v.i. ‘to look, to look at’, although these words are clearly in a polysemous, or more correctly heterosemous relation. In other similar cases, there is only one main entry—cf. hiku, kehe, kili², lima. The use of mata to mean ‘point, tip, blade etc of something’ is assigned to mata³.

Unfortunately, but not unexpectedly, this arrangement has presented the editors with another, more intractable problem: it is often difficult to assign derived forms to mata¹, mata², or mata³, and there are many inconsistencies. Why are mataola ‘to be healthy looking’, matamitaki ‘good looking’, and matakata, matakatakata ‘to have a smiling face, to look happy’ under mata²? I note that matahavala ‘mischievous, cheeky’, and matatetefu ‘stern-looking’—presumably compounds of the same type—are under mata¹, where I would expect them. A guiding principle (not a very good one, and not consistently followed) seems to be that mata¹ is mainly used for nominal forms and mata² for verbs, irrespective of whether the derived word or compound takes its meaning from the nominal or the verbal use of mata.

Mata³, as we would expect, contains mataihu ‘point of the nose’ and matavelo ‘sharp point of rock or land’. Why then are the following all listed as subentries under mata¹ when all are compounds of the type found under mata³: matafonua ‘headland’, matagao ‘very point of a corner, gable’, matahiku ‘extremity, tail end’, matakupega ‘point of a net’, matamui ‘end bit of something, tip’, and matapotu ‘end, endpoint’? Furthermore, matakeho kupega ‘the front point of the hand fishing net’ is in mata³—but at the end of the entry, out of alphabetical order, as though added as an afterthought. [end page 194]

Like the other dictionaries in the PALI series, this one has detailed user notes explaining the organization of the dictionary and the categories used in the notes and explanations. Part A is the Niuean–English portion of the dictionary. Section 1 is concerned with orthography and pronunciation, section 2 deals with the organization of entries, and section 3 contains a useful annotated bibliography with separate sections for flora, fauna, and grammar. Part B is the finderlist. Since these user notes raise a number of issues relevant to the dictionary itself, I will use them to structure my discussion.

Most of section 1 consists of a very clear explanation of the difference between long vowels, diphthongs, and rearticulated vowel sequences. Section 1.4 compares "New Zealand-English and Niuean alphabets and their phonetic realisations," combining an informal description of Niuean phonemes in terms of English correspondences (and occasional technical phonetic descriptions where the correspondences are not sufficiently close), with an explanation of the transliteration of English graphemes such as q, x, ph into Niuean. The question of intended audience becomes somewhat muddied here. As a reference resource, this section is serviceable enough, but it would have been more coherent if divided into two parts, one a description of Niuean phonemes, the other an account of how some English phonemes and written symbols are pronounced and written in Niuean. I have some other minor quibbles. English bread should have been contrasted with bet rather than bed to exemplify the difference between /ê/ [e-macron] and /e/; English ch as in cheque changes to Niuean si rather than s; and the transliteration of English consonant clusters into Niuean could have been explained by an actual example of a loanword, rather than a hypothetical one.

Section 2 takes pains to make transparent the process of constructing a dictionary entry. The entries contain the following information: part of speech designation; morphological composition of the headword if it is complex; grammatical notes where required, especially with grammatical morphemes; and the English definition or translation equivalent, including scientific names of species. Some entries contain additional semantic, sociolinguistic, or cultural notes, and almost every entry, except for species names and some borrowed nouns, is illustrated with example sentences. Source languages of borrowed words are given, and so are a certain number of Proto-Polynesian or lower-level reconstructions. (The latter are limited to cases where the etymon is listed in POLLEX [Biggs 1993] with a Niuean witness.) Finally, variant forms of words are listed and cross-referencing is used for suppletive forms and for some synonyms and antonyms.

These various categories of information are discussed systematically in turn, and the conventions for entries and subentries are explained. On the whole, I think this section does an excellent job of making the dictionary accessible to readers, including native speakers of Niuean, and that the level of technicality in the explanations is well-judged. If a version in Niuean could have been provided, this would of course have gone even further towards realizing the NLD’s stated aims. I particularly liked section 2.5 "Definition, meaning, gloss," which among other matters undertakes to explain the differing semantic ranges of seemingly equivalent words in two languages, and outlines the various strategies used to convey the meanings of Niuean words. [end page 195]

One of the most attractive features of the entries is the way in which the morphological composition of complex words (mostly subheadwords) is shown, for example:

fefakakiteaki, fe, faka, kite, aki (Prf., Prf., see, Suf.)

However, this procedure is largely limited to words whose morphology is transparent to present-day Niuean speakers, so its main usefulness would seem to be for learners of the language. Niueans themselves might have been better served if the current state of comparative and historical knowledge had been drawn upon to provide analyses of words deemed "lexicalised" and hence unanalyzable. However, this policy is explicitly eschewed (13).

Not all of the explanations contained in the front matter are entirely satisfactory, and I will now discuss some cases where the analysis adopted there has an adverse effect upon the dictionary entries themselves. One such area is the discussion of reduplication. NLD postulates three kinds of reduplication, as follows:

DUP stands for ‘complete’ reduplication, for example kalakala …
REdup stands for ‘partial’ reduplication, whereby only the first syllable is reduplicated, for example popoko …
reDUP stands for ‘partial’ reduplication, whereby only the last syllable [sic] is reduplicated, for example: mahalohalo … (14)

The problem lies with "reDUP." Some examples of words analyzed in this way in the dictionary entries are maafuafu ‘to perceive a smell in wafts’ [sic; a more accurate gloss would be ‘(of a smell) to be perceived in wafts’], maakiaki ‘to slip off’, maaliali ‘to be very clear’, mafâfâ ‘to be cracked open’, fakaatiati ‘to check on someone or something’, fakaoloolo ‘to reach out’. A description in terms of syllables is obviously unworkable (a rule in terms of two syllables would not suffice, because of examples like mafâfâ). This problem arises from the decision, referred to above, to make limited use of comparative-historical explanation. An historical process of full reduplication in addition to prefixation with ma- or faka- has been obscured by the loss of the simple root in Niuean. The concept of a bound root (cf. English -ject and -ceive) is not beyond the scope of NLD readers. Indeed, McEwen employs it, both in a brief comment at the end of his introduction (1970:xxiii–xxxiv) and in many entries. Lieber and Dikepa (1974) make "root" one of the organizing concepts of their Kapingamarangi lexicon. The difficulty with setting up a synchronic category "reDUP" is that it can only be described adequately in terms of an element that looks suspiciously like a prefix and another element that looks suspiciously like a stem. The analysis seems particularly infelicitous when the dictionary entry suggests a possible derivation in terms of proto Polynesian forms. Compare the first two entries below with the next two.

mafâ v.i. To be cracked open, to be split open … [PN *ma-faqa ‘broken, shattered …’] mafâfâ, reDUP. v.i. To be cracked open …

fakaati v.t. To check on someone or something … [PN *ati ‘keep a lookout for …] fakaatiati, reDUP … [end page 196]

lote v.t. To scratch … malotelote, ma, lote, lote (Prf., DUP). v.i. To be lacerated …

gega v.t. To cling … fakagegagega, faka, gega, gega (Prf., DUP). v.t. grammar: frequentative. To make someone climb up or cling to something …

Buse et al. (1996:viii) call similar cases in Cook Islands Mâori "last element reduplication" or "Rr." Examples are mâtorutoru ‘quite thick, firm and solid’, mâtautau ‘be accustomed, habitual’, mâteatea ‘open, unobstructed’. This solution begs the question of the morphemic status of the "elements," but at least avoids the solecism of using the term "syllable" for a bisyllabic element, and acknowledges implicitly that the ma-/mâ- element is the prefix of that form.

Another problematic area concerns verb categorization. In section 2.3.1.1, the major verb classes are described as follows:

A verb operating as ‘verb transitive’ is uniquely distinguished by having an agent-object ordered after the verb and before the unmarked (for case) patient-subject. On semantic grounds one can label such verbs also as ‘passive’, but with the proviso that there is no active-passive (or passive-active) transformation in Niuean. (17)

Intransitive verbs are distinguished by the fact that the unmarked (for case) actor-subject is the first nominal argument. On semantic grounds such verbs can be labelled active or stative … (18)

The terminology used here seems unnecessarily confusing. "Patient-subject" has a certain currency and has been familiar in Polynesian linguistics since Biggs (1974). However the term "agent-object" is not to my knowledge used elsewhere and has a paradoxical ring to it. It is not clear what the "semantic grounds" are for calling transitive clauses passive. It seems to me that the grounds are formal, and suggested rather by an analogy with the prepositional marking of agent in English passive. A Niuean transitive clause without an expressed agent requires a passive translation in English, and this should be pointed out, but to say that a transitive clause with an overt agent and patient is a passive construction is another matter entirely, and not a solution usually chosen in grammars of Polynesian languages. (For one thing, it passes over any consideration of the pragmatic or discourse function of transitive clauses in Niuean as compared to passive clauses in languages with active–passive alternations.)

The decision to avoid any mention of ergativity was probably a sensible one, given the predominantly lay readership of dictionaries, but it need not have given rise to the anomalies described above. Neither McEwen in the introduction to his Niue dictionary nor Churchward in his Tongan grammar (Churchward 1953, chaps. 10 & 11) uses the term ergative, and their comparatively nontechnical descriptions are comprehensible to any reader who understands the traditional grammatical categories of transitive, intransitive, subject, and object.

Although the editors decided to dispense with the term "ergative" in describing the argument structure and case-marking of Niuean, the expression "in the ergative analysis" crops up in the dictionary glosses for the agentive case-markers he and e and the common article e. These explanations will baffle the reader who has attempted to master the terminology presented in the user notes. Moreover, it is disconcerting to see the expression "in the ergative analysis" used both here and elsewhere (Sperlich 1994) in a way that suggests that this is a question of some exotic and controversial theory about Niuean, rather than a frequently encountered and easily describable case-marking pattern.

The shortcomings in the description of the verb classes need not have affected the dictionary proper, but unfortunately the straightforward distinction based on case-marking of core arguments is not applied consistently in the entries, where many syntactically intransitive verbs (as clearly indicated by the syntax of the example sentences) are labelled v.t., perhaps because they seem to be semantically transitive, for example, nava ‘to praise, to admire’, and paleko ‘to issue a challenge’. The verb mau ‘to be complete, to be finished, to be held’, with a subsense glossed ‘to marry’, is labelled v.t., although none of the example sentences contain agent and patient arguments appropriately marked, and cognates of mau in other Polynesian languages are intransitive. The reduplicated form kitekite is also marked as v.t., although the examples demonstrate that it is transitive when it means ‘inspect’ but intransitive when it means ‘to look carefully, observe’.

The problem seems to be particularly acute in the case of faka- derivatives. We are told, "Niuean verbs prefixed with the causative faka- are all transitive, as there is a cause (agent-object) expressed or implied" (18). Polynesianists will regard this statement with suspicion. In fact, in Niuean as elsewhere, in addition to true causatives such as fakalologo v.t. ‘to make someone sing’, there are many faka- derivatives that are intransitive. This dictum is not consistently applied in the dictionary itself, where many faka- verbs are quite properly labelled intransitive—although unfortunately other clearly intransitive verbs are labelled transitive! So we find fakataane v.i. ‘to behave like a man’, and fakafifine v.t. ‘to make out to be like a woman, to be effeminate’. The textual example accompanying fakafifine contains an intransitive construction with one participant. Many faka- derivatives are made to appear transitive by being given a reflexive gloss, as in fakamatatetefu ‘to make oneself look sternly’. But once again the example sentence has the give-away intransitive case-marking pattern.

Some of these miscategorizations may be typographical errors, as not even the glosses suggest transitivity: magihogiho ‘to be very itchy, to be very hot (of spices)’, fakafôhake ‘to [lie] face up’, fakafôhifo ‘to [lie] face down’, fakanonâ ‘to stop beating (of heart)’, and faite ‘to pose, to expose oneself’.

Reciprocal verbs are also handled inconsistently. These take a plural subject in the unmarked case, but fekiteaki ‘to see each other, to meet’ is classed v.t., whereas felogonaki ‘to hear each other’, is correctly classed v.i.

As noted above, the term "active" is introduced as a semantic class of intransitive verb, in contrast to stative (18). However, the term occurs again in a context that is incompatible with this definition. In the dictionary entry for the suffix -i we are told that it can change passive/agentive type verbs (i.e., transitive verbs in the terminology used in the introduction) into "active type" verbs. However in the only example given, maga, maga-i, the suffix changes a stative intransitive verb into an agentive one—in line with the use of the same suffix in Tongan, but quite contrary to the explanation given in the grammatical note. Since the distinction between [end page 198] active and stative (or more usefully between actor-subject and patient-subject) intransitive verbs has no material consequences for the dictionary, it need not have been introduced.

Some other grammatical terms are used inappropriately. I will discuss two more instances. In the entry for the prefix ma- we are told that it is "attached to verbs to transform them into participles, often indicating some natural or accidental happening, agent unknown or irrelevant" (193). The semantic part of this note is correct, but the use of the term participle is an error that can be traced back to a number of 19th-century descriptions of Polynesian languages, and that probably arose because ma- derivatives, which have a semantic patient in the unmarked case, require passive-voice English translation equivalents. The ma-derived verb in the example sentence given here, and also those in the examples for all the other ma- derivatives I looked up, are finite verbs with tense-aspect marking and no constituent that can properly be called a participle. (The phrase "a verbal complement sentence" in the entry for ma- also presumably arises from confusion with the English translation equivalent.)

The verb mao ‘to be perceived’ is said to occur with an incorporated object, but in fact the compounds to which the grammatical note refers, such as mao elo ‘to be perceived as stinking’ and mao vela ‘to be perceived as hot’ [my glosses] are formed with a stative verb or adjective. The terminology is confusing and inaccurate, since incorporated-object constructions, in the normally understood use of the term, also occur in Niuean, for example, unu mena [wash (of clothes) + thing] ‘to wash clothes’ (346), ati tupe [build + money] ‘fundraising’ (52).

Ironically, most of these errors and confusions arise in areas where the editors have attempted to provide grammatical information beyond what is necessary for strictly lexicographical purposes. They will be of small concern to most of the Niuean speakers who consult this volume, but they will be a hindrance to any who seek to enlarge their understanding of Niuean grammar. To linguists they are an irritant, and to learners of Niuean they are a potential source of misinformation.

The computer-generated English–Niuean finderlist occupies over 230 pages, and is explained in part B of the user notes. The editors have been generous in the flagging of keywords, and many entries contain comprehensive lists of phrases, idioms, and compounds using the headword. The question of intended use always seems to me particularly ambiguous in the case of finderlists. This one is fascinating in its abundance, but would be of most use to an English speaker or linguist attempting to learn Niuean, or possibly to someone making translations from English into Niuean. These considerations apart, I have just one minor criticism of this section, concerning the organization of entries. Only in cases where a single keyword corresponds to a headword is the reverse English headword followed directly by the Niuean gloss. Owing to the rigorous application of this principal, there are many entries (e.g., for eye, run, voice, woman) where one has to pass through many phrasal subentries in alphabetical order before coming to the most obvious translation equivalent of the headword. Of course this will not hold up a practiced user for long, but it can be an annoyance nevertheless.

The text is remarkably clear of typographical errors, apart from a number in the user notes. One that might cause misunderstanding occurs in the sentence, "Local [end page 199] nouns generally behave like ‘proper names’ in that they do not occur with the proper article" (19), where the word in should presumably be replaced with except.

The foregoing criticisms are offered in a constructive spirit and the hope that a future edition of the NLD may be able to correct some of the inconsistencies. In the most essential aspects of its content, this is a splendid book; I just wish it were perfect.

Biggs, Bruce. 1974. Some problems of Polynesian grammar. Journal of the Polynesian Society 83:401–426.

———. 1993. POLLEX: The Comparative Polynesian Lexicon Project. Computer file, University of Auckland.

Buse, Jasper, with Raututi Taringa, ed. by Bruce Biggs and Rangi Moeka‘a. 1996. Cook Islands Maori dictionary with English finderlist. Pacific Linguistics C-123. Canberra: Australian National University.

Churchward, C. M. 1953. Tongan grammar. London: Oxford University Press.

Lieber, Michael D., and Kalio H. Dikepa. 1974. Kapingamarangi lexicon. Honolulu: University of Hawai‘i Press.

McEwen, J. M. 1970. Niue dictionary. Wellington: Department of Maori and Island Affairs.

Pukui, Mary K., and Samuel H. Elbert, eds. 1986. Hawaiian dictionary, rev. & enl. ed. Honolulu: University of Hawai‘i Press.

Sperlich, Wolfgang B. 1994. A theory of verb classes and case morphology in Niuean. Paper presented at the Seventh International Conference on Austronesian Linguistics, Leiden, 22–27 August.