Halkomelem

Halkomelem (;[3] Halq̓eméylem in the Upriver dialect, Hul̓q̓umín̓um̓ in the Island dialect, and hən̓q̓əmin̓əm̓ in the Downriver dialect)[4][5] is a language of various First Nations peoples of the British Columbia Coast. It is spoken in what is now British Columbia, ranging from southeastern Vancouver Island from the west shore of Saanich Inlet northward beyond Gabriola Island and Nanaimo to Nanoose Bay and including the Lower Mainland from the Fraser River Delta upriver to Harrison Lake and the lower boundary of the Fraser Canyon.

In the classification of Salishan languages, Halkomelem is a member of the Central Salish branch. There are four other branches of the family: Tsamosan, Interior Salish, Bella Coola, and Tillamook. Speakers of the Central and Tsamosan languages are often identified in ethnographic literature as "Coast Salish".

The word Halkomelem is an anglicization for the language Hul'qumi'num, which has three distinct dialect groups:

The language differences (namely, in phonology and lexicon) are greatest between the Island and Upriver dialects, with the Downriver dialect (especially the Tsawwassen First Nation) providing a central link between the other two. The diversity of the Halkomelem dialects is noted to be the result of complex social and economic forces and linguistic change, as many Island people crossed the Georgia Strait to camp along the Fraser River (in both the Downriver and Upriver areas) for the summer runs of salmon. Arranged marriages between children in different language areas was also common, helping to establish a regional social network in the Strait of Georgia–Puget Sound Basin.

The Halkomelem language is near extinction. In 2000, it was estimated that the number of fluent Halkomelem speakers was less than twenty-five. Most are middle-aged or older, and few are monolingual, as there was a flood of English-speaking settlers in the region in the mid-19th century. Language programs at the Stó:lō Nation, Seabird Island First Nation, and Cowichan First Nation have been developed to save the language. A program aimed at adults at Musqueam is a collaboration between the band and the University of British Columbia First Nations and endangered languages program.

In September 2009, the University of California Press published linguist Brent Galloway's Dictionary of Upriver Halkomelem.[9]

A Halkomelem iPhone app was released in 2011,[10] and an Android version was released in 2016.[11] The app was developed by the FirstVoices website. There are 1754 words archived and 690 phrases archived on the FirstVoices website.[12] As of 2014, 263 fluent speakers had been reported.[13] In 2014, the number of Head Start Programs was 21, and this included a language-nest immersion preschool.[14]

Note: All examples are drawn from the Downriver dialect of Halkomelem spoken by the Musqueam band. Relevant differences in the phonology of the Island and Upriver dialects are noted at the foot of the phoneme charts.[15]

Halkomelem has five vowel phonemes. Long and short vowels (but not schwa) contrast. Vowel length is written in the native orthography as ⟨·⟩.

All five vowel phonemes vary considerably phonetically. The phoneme /i/ has three distinct allophones. It is realized as [e] following unrounded uvulars. It is realized as [ɪ] with a central off-glide preceding both unrounded and rounded uvulars. Elsewhere, it is realized as low [i] or high [e]. The /e/ is realized as a low to mid-front vowel, usually between [ɛ] or high [æ]. The /a/ is low and central to back, often close to [ɑ]. The /u/ is high, back, and rounded, realized somewhere between low [u] or high [o].

When stressed, the schwa /ə/ appears in most environments as a mid-central, but it is fronted and raised before /x/, approaching [ɪ]; before /j/ it is also fronted, approaching [ɛ]; before /w/ it is lower and back, approaching [ɑ]; and before rounded velars it is mid-back, close to [o]. Unstressed /ə/ can be as high as [ɪ] before /x/ and /j/, and before labialized velars it is realized as [o] or [ʊ]. This phoneme can also be assimilated to a stressed /e/ or a stressed /a/ in an adjacent syllable, by vowel harmony.

The plain plosives are less aspirate before vowels than in English, but they are more aspirate finally. Although the glottalized plosives are ejectives, they are not usually strongly released.

Suttles (2004) makes several interesting notes on the Musqueam obstruents. The labiodental fricative /f/ occurs in recent loans from English and their derivatives such as in káfi "coffee" and in číf "chief." The stops /t/ and /tʼ/ are articulated at a point slightly forward of that of the usual English /t d/, while the affricates c /ts/ and /tsʼ/ are somewhat more retracted than these same English /t d/. The affricate [d͡ʒ] has only been recorded in kinjáj "English people" and kinjájqən "English (language)." The glottalized lateral affricate /ƛʼ/ [tɬʼ] is produced when the apex of the tongue at the onset is in the position for the lateral release rather than for a /t/, and there is less friction produced than with other affricates. The phonemes /k/ and /kʼ/ occur in "baby talk" as substitutes for /q/ and /qʼ/. The uvular fricative [χ] is produced with a great deal of friction and/or uvular vibration, and it contrasts strongly with the velar fricative /x/.

There is variation in the extent to which Musqueam speakers glottalize resonants. Phonetically, there are glottalized resonants (e.g. [nˀ]) and resonants preceded or followed by glottal stops (e.g. [ʔn nʔ]), however, Suttles (2004) finds no instances of contrastive distribution among any of the three. He puts forth two explanations for these facts: that there are two sequences of phonemes, /Rʔ/ and /ʔR/, with overlapping [Rˀ] allophones, or that there is a single phoneme /Rˀ/ that is realized in three distinct ways. In preferring the latter explanation, Suttles holds that there may be five glottalized resonant phonemes in the dialect, although Downriver speakers glottalize resonants very lightly, making them difficult to detect.

In most Upriver dialects, glottalized resonants do not exist, while in Island dialects, they are more sharply articulated (tenseness is a key feature of Island speech). As is the case with many other phonological features, Downriver Halkomelem stands as a link between the other dialect areas, and it is possible that its speakers vary depending on Island or Upriver influence. Other differences between dialects include: Island and Downriver have both /n/ and /l/, while Upriver has merged these as /l/. Upriver Halkomelem lacks the post-vocalic glottal stops of the other two dialects, and shows compensatory lengthening in that environment. Additionally, Upriver dialects have greater pitch differences, and some words are differentiated by pitch alone.

Based on Suttles' (2004) recordings of several speakers of the Downriver (Musqueam) dialect, stress in Halkomelem consists of an increase in intensity and an accompanying rise in pitch. The three levels of stress are primary (marked /׳/), secondary (marked /`/), and weak (unmarked). There is one vowel with primary stress in every full word, however, its occurrence is not completely predictable.

In uninflected words with more than one vowel, the primary stress usually falls on the first vowel (e.g. as in céləx "hand" and léləmʼ "house"). There are exceptions to this general pattern (e.g. as in xəmén "enemy"). As shown by the preceding example, if the word contains both a full vowel and one or more schwas, the stress is placed on the full vowel. Again, there are exceptions to this pattern, such as in words with a final glottal stop that cannot be preceded by schwa (e.g. as in nə́cʼaʔ "one").

Although minimal pairs contrasting stress are rare, they do exist in the language. The primary stress of a verb root consisting of a resonant, a schwa, and an obstruent followed by the suffix /-t/ "transitive" can fall on either the root or the suffix, allowing for minimal pairs such as mə̀kʼʷət "salvage it" and məkʼʷə́t "finish it all."

The secondary stress appears most often in words that are composed of a root that has retained its stress and a stressed suffix (e.g. as in cʼéwəθàmx "help me"). It may be the case, however, that the secondary stress recorded by Suttles (2004) in words like cʼéwəθàmx is actually a falling pitch; this seems to be characteristic of the last stressed syllable of a phrase in the language. Additional analyses of the sentential intonation patterns are needed.

All obstruents (except the glottals) typically follow one another in sequences of up to four, although a sequence of five is also possible (e.g. as in txʷstx̌ʷásʔal "just standing in shock"). There are no specific restrictions on the types of obstruent sequences that can occur. Plosives appearing in sequences are rearticulated, and sequences of /ss/ are common in the language.

Resonants only appear adjacent to vowels. When these sounds occur in the middle of words, they are found in sequences of resonant-obstruent, resonant-resonant, and obstruent-resonant. An initial resonant is always followed a vowel, and a final resonant must be preceded by one.

The laryngeals are more restricted than members of the other natural classes in Halkomelem. The glottal stop occurs only adjacent to a vowel, and, within words, it does not follow any obstruent except (the prefix) /s/. It can never occur in final position following a schwa. /h/ occurs only before vowels, following a resonant or one of the fricatives at morpheme boundaries, but never following other obstruents. It can appear between an unstressed and a stressed vowel, but it cannot occur between a stressed and an unstressed vowel.

Certain processes affect the realization of underlying sounds in Halkomelem. Alternations that occur fairly commonly are discussed in this section, rather than in the following section on morphology.

In 1997, the Musqueam First Nation officially adopted the Americanist phonetic alphabet.[18] This alphabet does not use upper case letters.

Like the majority of Salishan languages, Halkomelem is polysynthetic. A word in Halkomelem may consist of a root standing alone and unaltered, or of a root altered by one or more processes of internal modification and/or accompanied by one or more affixes. Since all words (with the exception of a few adverbs) can function as predicate heads, there is no basis for distinguishing verbs, nouns, and adjectives. There are other bases, however, for distinguishing these classes. Verbs have progressive forms and do not take possessive affixes, while nouns do not have progressive forms and do take possessive affixes. Adjectives have neither progressive forms, nor do they take possessive affixes. Compounding is non-existent in the language, although some scholars believe to have found a few possible examples.[20]

The majority of verb roots have the shapes CAC, CəC, CəCC, while noun roots typically have the shape CVCVC (V is any vowel). The most common shapes of adjective roots are CəC and CAC. There is a prefix that nominalizes verbs and adjectives, and there are several prefixes that make verbs out of nouns. Additionally, there are several ways to make adjective-like words from nouns. Processes of internal modification of the root include reduplication (of initial CV and CVC), shift in stress and vowel grade, and glottalization of resonants (which also affects suffixes). Roots of different shapes often undergo different processes to produce forms that are grammatically identical.[20]

Verbs roots are identified as perfective, as opposed to progressive, aspect.

Several verbs also have a durative aspect, which can occur in both forms.

A number also have an iterative-dispositional aspect. For a few of these verb roots, this aspect can appear in both a progressive and in a perfective form.

The majority of verbs have a resultative form which is adjective-like and does not carry a progressive-perfective distinction.

The plural can be optionally marked in all of these forms. The diminutive is also marked, optionally, in only the progressive and resultative aspects.

It is possible to internally modify noun roots in Halkomelem for the plural, the diminutive, and the diminutive plural. Compare:

A few nouns may have resultative forms. They do not have progressive forms, but they may be made into a verb with a verbalizing affix and then express this form.

Similar to noun roots, adjective roots can be internally modified for plural, diminutive, and diminutive plural. They can only have progressive forms if made into verbs by means of a verbalizing affix.

Complex adjectives are formed from adjective roots and lexical suffixes.

Halkomelem contains prefixes, suffixes, and infixes. All infixes of the language have been described in the preceding sections. Affixes are typically divisible into inflectional or derivational and grammatical or lexical categories, depending on their involvement in paradigms and meaning, however, a number of Halkomelem affixes mix these categories. Suttles (2004) identifies the following classes of suffixes and prefixes; a sampling of these affixes follow.

The following table lists the possessive affixes which appear in attributive possessive structures in Halkomelem.[21]

Possession is marked either on the possessed noun (the head) or the word preceding it through these affixes. Together with the appearance of affixes, possession also requires a structural component, in that the possessor of the head is found to the right of the head. The possessor is always preceded by a determiner, although depending on the noun class, it can also appear with an oblique case marker. If the possessor is a common noun, it will be introduced by a determiner, but without an oblique case marker. If, however, the possessor is a proper noun, it must appear in the oblique case. Thus, it will be preceded by an oblique case marker, and the possessed noun will appear without a possessive affix. For proper nouns, the determiner and the oblique case marker are fused into a single particle. Marking common nouns with an oblique case marker results in an ungrammatical construction:[22]

Most verbs roots are semantically patient-oriented (e.g. they have glosses like "get hit" or "get washed"), while few verbs are semantically agent-oriented (e.g. "look" or "see"). All are grammatically intransitive. These relations are different with the suffixes of the voice system. A verb that is made up of an inactive root and an intransitive suffix is grammatically intransitive, but semantically active. An inactive or active root that takes on a transitive suffix is grammatically transitive and takes an object. The transitive suffix is the base for an object or passive person suffix. Two of the most commonly used transitive suffixes distinguish actions performed with limited control or accidentally from those performed with full control or purposely.

Aspectual prefixes, which precede predicate heads, have adverbial meaning and express temporal distinctions. Modal suffixes follow the suffixes of the voice system and indicate desire or intention and search or arrangement.

Lexical suffixes can be related to verb roots as objects, locus, or instruments; to adjective roots as noun heads; and to noun roots as noun possessors or the noun heads of modifiers.

The personal affixes distinguish first, second, and third person in singular and plural. There are neither dual forms nor inclusive/exclusive distinctions in this language. There is also a set of possessive affixes (prefixes for first and second person singular, suffixes for first-person plural and third person, and a combination of prefix and suffix for second-person plural). This system will be covered, in detail, in the "Syntax" section.

Derivational prefixes and suffixes form an inner layer around the word root, while inflectional affixes form an outer layer around the root. Among derivational affixes, those with lexical meaning stand closer to the root than those with purely grammatical meaning. Among inflectional affixes, those of the voice and person systems stand closer to the root than the aspectual prefixes and modal suffixes.

In Musqueam, a sentence minimally consists of a predicate. Predicate heads can be bare roots (e.g. cákʼʷ "far"), derived forms (e.g. spéʔeθ "black bear"), inflected forms (e.g. cʼéwət "help him/her/them"), and forms including both derivational and inflectional affixes (e.g. kʷə́xnəct "name-base-transitive," as in "name a price"). Predicate heads can also be words that are definable morphologically. This includes verbs (e.g. ném "go"), adjectives (e.g. θí "big"), nouns (e.g. swə́yʼqeʔ "man"), members of the closed sets of personal words (see the following section), and interrogative words (e.g. stém "what").[23]

A verbal predicate may be expanded with the addition of one or two auxiliary verbs before the head and/or one or more adverbs preceding or following the head.

Like verbal predicates, adjectival and nominal predicates can be expanded with other elements. Nominal predicate heads can appear alone, followed by particles and adverbs.

Syntactically, adjectival and nominal predicate heads appear with auxiliaries less often than verbs. Adjectives and nouns can appear together in nominal predicates expanded with adjectives. A nominal predicate head can be preceded by a modifying adjective or numeral.

Adjectives usually appear as predicate heads accompanied by particles only, but they can be preceded by auxiliaries and adverbs acting as intensifiers.

The Halkomelem person markers (forms that correspond in meaning to English personal pronouns) include a set of affixes, one set of particles, and two sets of words (personal and possessive). As mentioned in the "Morphology" section, there is no dual number or inclusive/exclusive distinction in the language, however, some scholars believe that the forms identified here as second-person "singular" were once used in addressing a married couple, a pair of brothers, or even a family, while the "plural" forms were used for a larger or less integrated group.[24]

The first- and second-person particles pattern like a nominative–accusative case marking system. In other words, the same particles mark first- and second-person arguments in both intransitive and transitive predicates in main clauses (coordinate constructions).

Suttles (2004) classified the first- and second-person argument particles as second-position predicate particles, along with about twenty other particles that can appear within the predicate. Most of the second-position predicate particles are mobile, appearing after the first word of the predicate (whatever that may be). If the only word in the predicate is the head, the first- and second-person argument particles will follow it; if the head is preceded by an auxiliary verb, they will follow the auxiliary; if the head is preceded by an adverb, they will follow the adverb.

An object person marker can only be suffixed to a transitive verb (i.e. a verb stem that has a transitive suffix). The four forms appear to be composed of identifiable elements: /-S/ and /-am/ "non-third-person singular," /-al-/ "non-third-person plural," /-x/ and /-xʷ/ "first person," and /-ə/ second person. Third-person objects are unmarked.

The element /-S/ occurs with the transitivizer /-t/, and they coalesce as /θ/. With the root cʼéw- "help" and /-t/ "transitive," we find:

The third-person arguments follow an ergative–absolutive system. With an intransitive predicate head, a third-person argument is like a third-person object in being marked by zero. Plurality is optionally indicated by the particle ʔé·ɬtən.

With a transitive predicate head in a main clause, on the other hand, a third-person agent must be marked by the suffix /-əs/. It always follows the transitivizer and object person marker, if any. Unlike the first- and second-person particles, the suffix does not move to follow an auxiliary or adverb. Again, the plurality of the third-person may be indicated by the particle ʔé·ɬtən.

In the active paradigm, a third person cannot be the agent with a second person as the object. Instead, we find passive forms.

Only third persons can be agents in the passive. Other relations (e.g. the forms "*I am seen by you" or "*he is seen by me") can be expressed only in the active (e.g. "you see me" and "I see him").

A subordinate clause is produced by prefixing one of the two subordinating particles, /wə-/ "if, when, that" and /ʔəl/ "whenever, whatever," to the first word in a predicate and replacing its coordinate agent marker with a subordinate agent marker. Subordinate agent markers are the same in both intransitive and transitive active predicates.

Subordinate clauses usually follow main clauses, but there are a few exceptions.

"If people grow up acquiring more than one language, their ability to acquire additional languages is much greater," Kesler explains. "So when parents are thinking, 'Would I really rather my child learning Halkomelem than French?', yeah, I think that's not a bad idea at all."[26]

— 
Linc Kesler, University of British Columbia associate professor and director of the First Nations Studies Program and First Nations House of Learning