Research summaries

My goal is to share my research so that it is accessible to people without linguistic training. I welcome community comment/feedback to make these summaries better -- do not hesitate to email me. :)

For full citations and links to the papers themselves, see my Publications page.


Personal and traditional stories

We have recorded thirty-six stories told by the speakers we work with regularly in Vancouver: Barbara Sennott (Xsim Mihl'mihl), Vince Gogag (Gawagani), and Hector Hill (Sgaak). The stories include personal anecdotes, recipes, discussions on parts of Gitxsan language and culture, and traditional histories. Our goal is to produce written Gitxsanimx̲ versions of all of these stories alongside translations and linguistic analysis of each individual word. We also aim to discuss some of the unique properties of each speaker's dialect and show that dialect differences are much more detailed than the general idea of "Eastern" and "Western" -- some differences only involve pronunciation, others involve, for example, differences in how to convey emphasis on what is being talked about, or ways of connecting sentences together in extended speech.

Agreement and pronouns

My work on agreement explores how pronouns (words like I, us, you, him, her, them) appear in various types of Gitksan sentences. In Gitksan, there are three categories of pronouns: free pronouns ('nii'y, 'nuu'm, 'niin, 'nit, 'nidiit), suffixes (-'y, -'m, -n, -t, -diit), and a special group of floating elements called clitics near the front of a sentence (n, dip, m, t). The pattern of how to use these three types of pronouns accurately in Gitksan is very different from how it is done in English and most other major world languages, and so usually does not come very naturally for a learner who is only familiar with English. First, it is important to know the type of sentence: a sentence that begins with nee(dii) 'not' or yukw '-ing' or wil/win 'that' uses the suffixes by default, while a sentence which does not have those words will use the free pronouns by default. For example, the me/I pronoun is different in a plain sentence compared to a sentence with needii.

  1. Bax̲ 'nii'y. 'I ran/I run.'

  2. Needii bax̲a'y, 'I didn't/don't run.'

Second, the subject of a sentence that also has an object will always use a special pronoun. Subjects with this status are called the ergative subject. In a plain sentence, the ergative subject is marked with a suffix pronoun while a sentence dependent on nee(dii) or yukw will have the ergative subject marked with a clitic before the verb.

  1. Gya'a'y 'niin. 'I saw/see you.'

  2. Neediin gya'an. 'I didn't/don't see you.'

These facts were documented about Gitksan decades ago, and they are also largely true about the neighboring Tsimshian and Nisg̲a'a languages. However, the exact pattern is very unusual and hasn't been documented in any other group of known languages. This makes them a very interesting set of facts to help us better understand how pronoun systems can work in human language more generally. I have been working specifically on sentences involving the pronoun 'nidiit/-diit 'them', which breaks some of these rules.

Word structure: prefixes, suffixes, and roots

The pronunciation of a Gitksan word can help you understand more about its structure/meaning, and vice versa. The aspect of pronunciation that contributes to this is where the "stress" (or loudest syllable), comes inside a word. In English, stress can be used to distinguish some words like próduce (fruits and veggies) and prodúce (make or generate something). In Gitksan, stress usually appears in a consistent place in a word: on the root, avoiding any prefixes or suffixes. For example, the two words kw'ootxw 'be missing' and kw'oodin 'lose' are related because they have the same root, and the emphasis in both words falls on that root, *kw'oo(t). Other words are related because they share an old suffix. For example, algyax̲ 'speak', gwalgwax̲ 'thirsty', and even Gitxsanimx̲ 'Gitxsan language' all share a suffix -(a)x̲ that descended from the word aak̲ 'mouth', though it doesn't literally mean this anymore. This -ax̲ suffix does not have any emphasis.

In the majority of cases the root has only one vowel, once we take all these prefixes and suffixes away, and the stress on a word is found there. For a small number of roots, there is more than one vowel to choose from, such as gwila 'blanket' or naasik' 'raspberry'. In these cases, the stress is usually on the last long vowel (náasik'), or if there is no long vowel, then the last vowel (gwilá).

Other topics in grammar

Plurals. Gitksan has many different ways of marking plurals! Unlike English, which only marks plurals of nouns (for example, cat/cats), Gitksan marks plurals on nouns but also, more often, verbs (hats'/hashats' 'bite many things, bite many times'). For names and some special nouns, there is also a special plural marker dip that points out that the person you are talking about is not alone, but instead is acting with a group. For example, you can talk about dip Clarissa 'Clarissa and the person or people with her' for a useful way of talking about what I did with my friends, or what I did with my family, without having to say specifically who I was with. The person you are talking to will usually understand from context. This is a group-marking plural that is completely missing in English, but common in Gitksan, and very useful in conversation.

Coordination and conjunction. The Gitksan word g̲an 'and', which links together nouns (like in nox̲o'y g̲ant nigwoodi'y 'my mom and dad'), has some special properties. Unlike in English, where the two nouns stay side by side around the word 'and', in Gitksan g̲an and the second noun can float away to the end of the sentence. I suggest that this construction is more similar to the grammatical structure for the English word 'with', and is similar to conjunction words in Slavic languages like Polish that have been translated as 'and' or 'with'.

Categories. It has been thought that all languages have the different categories noun and verb. It is a bit more controversial whether all languages also have a separate category of adjectives. I have explored whether Gitksan separates all three of these categories. In Gitksan, words like bax̲ 'run' (maybe a verb) and ts'uusx 'be small' (maybe an adjective) behave very similarly, so it is hard to tell at first glance whether they are both verbs, or if they are a verb and an adjective. I point out a specific area where they do not behave the same: when modifying a noun, verb-like and adjective-like words can appear in different places. Verb-based modifiers come after the noun, like in smax bahat 'bear that is running'. Adjective modifiers can come before the noun, like in t'uuts'xwit smax 'bear that is black'.