Hiding Waters reference grammar

Preface

Typically, reference grammars open with some kind of statement about who the grammar is for. "For the community of fluent speakers", it might say. "For teachers of the language." "For Serious Scholars of whatever language family."

If those are the usual audiences served by a reference grammar, it serves to highlight the silliness of this one, as the language described here is a constructed language. It is quite thoroughly made up. Its only speakers are fictional, its only scholar is me.

(I should note, though, that I do find it a little vindicating that even Real Linguists writing Real Reference Grammars feel compelled to spend their introductions awkwardly trying to justify what in God's name they did all this for.)

People who make up languages for fun are no strangers to the question, asked in equal parts by perplexed onlookers and by themselves, of why they are doing this. There is seldom any answer beyond "I dunno. Why do painters paint?"

Creating a language is delightful and surprising: some parts engineering, some parts treasure-hunting, some parts walking around outside until the trees slip you the answer. I may not have any scholars or speakers to write this grammar for, but to a great degree, I am writing it for myself. Trying to get the language out of my head and onto paper is one method of organizing the discoveries I've already made, and in some ways an act of further exploration, as it reveals gaps I have yet to fill.

Still, to say this grammar is an entirely personal matter would be a bit disingenuous. After all, if I am the only intended audience, then what am I doing posting it here on the public internet?

I suppose this grammar is meant for, if anyone, the broader conlanging community. Every project benefits from criticism and fresh pairs of eyes. (And obviously, any time someone looks at my work for long enough to offer feedback, I melt into a little puddle of delighted goo, regardless of what they actually say). Perhaps an idea I have described here will inspire someone to make something entirely new, and the world will be better for that too.

Another thing linguists do in the introductions to their reference grammars is write a list of thank-yous to the people who made the work possible. I may not have a university linguistics department who funded me, no professor or advisor who supervised the work. But I do have something else.

I have all the authors from the Domhantir project, years ago, whose encouragement got this whole show off the ground. I have Jim Henry, who very generously took the time to write a whole review of the language once, back when it was going by "Feayran" (meager as it was at the time). I have the folks at the CBB and ZBB whose feedback and criticism helped shape the thing—Micamo, Eldin, Ossicone, Astraios, and especially Rickard, whose enthusiasm never failed to make me feel like a million dollars. Countless others. Thank you all.

So. Here it is. May it bring you delight, as it brings to me.

Introduction

For the conlang connoisseurs and typologists in the audience, Hiding Waters is an artlang with some engelang habits that it should probably be more embarrassed by than it is.

An artlang, in conlanger jargon—or more precisely, the fictional language subtype of artlang, which is the most common one—is a language invented to be spoken by fictional people in some kind of fictional setting. Artlangs tend to aim for naturalism, whatever "naturalism" means where they're from. Tolkien's languages were artlangs.

Engelangs, on the other hand, are not particularly interested in naturalism, but instead aim to explore some kind of philosophical question or technical challenge. John Quijada's Ithkuil, which explores the question of how a language might achieve "the highest possible degree of logic, efficiency, detail, and accuracy in cognitive expression", is an engelang.

As is usually the case with any kind of typological system, these categories don't rigidly constrain the sorts of languages people create. You might be tempted to say, "Sure, it's more of a spectrum," but even that is wrong. Artlang and Engelang aren't the two extremes of a linear universe—they're just two of many peaks in a vast and varied landscape. Any given language likely falls somewhere off the slopes of the great mountain ranges, and not on the theoretical summit of any of them.

In the case of Hiding Waters, it lives fairly near the treeline of the Artlang Range. It was designed for a fictional group of people in a fantasy world, and tries to be an evocative lens with which to view the culture of its imaginary speakers.

However, it also invests in exploring some rather engelang-y questions:

• If there was a kind of people that thought primarily in smells and only secondarily in images, how might that affect their language?
• How might the natural ability to shapeshift affect how people thought and talked about the world?

Hiding Waters pursues these questions in particular because the way a language can teach you new ways of conceptualizing the world is what intrigues me most when learning and creating languages. Thus, Hiding Waters tries to be a radical departure from the modes of thinking I'm accustomed to in the languages I use day to day.

This is also one of the things holding Hiding Waters back from the peak of the Artlang mountains. Artlangs, in their quest to be naturalistic products of their fictional environments, tend to be very invested in the historical processes which shaped them into their present state, and all the grammatical and phonological artifacts such processes leave behind.

But while I don't find diachronics and phonology un-interesting, they tend to get out-competed in the grand struggle for my attention by things like interesting syntactic structures and cool extensions of novel conceptual metaphors. I have some ideas about a diachronic account of Hiding Waters, and some deeper phonetic details...which I'm sure I'll get to eventually. But for now, you'll find any mention of diachronics to be quite missing from this document, and the Phonology section to be about as minimal a sketch as one could conceivably get away with.

In exchange, I hope to provide some interesting explorations of the deep wilderness of register, anaphora, and morphosyntactic possibilities.

Points of Interest

If you're an itinerant conlanger passing by on the internet, you want to know what parts of this drawn-out document are worth your time. Where's the good stuff? If you'd like my entirely unbiased suggestions, these are the parts of the language I think are most unique or otherwise interesting:

The concept of stance is Hiding Waters' most defining characteristic, and one of the few features that have stuck around since the very beginning. It is something like a system of asymmetric formality registers, but it has some delightful twists—I especially like the interesting things stance can do involving discourse dynamics.

As far as I can tell, Hiding Waters lacks a morphosyntactic noun/verb distinction. That is, in any given sentence, there isn't a meaningful way to differentiate between constituents that are nouns and constituents that are verbs. Instead, there are predicates, which all exhibit uniform morphosyntactic behavior. (If you'd like to try your hand at morphosyntactic analysis, this is one area where I'd especially love some more perspectives!)

While the small number of morphosyntactic parts of speech limits the variety of syntactic structures there are to play with, the language still manages to be expressive via some clever systems of anaphora. Part of that apparatus is the way the language uses classifiers to create, maintain, and redirect references, which is delightfully fun—somewhat like a spoken version of American Sign Language's system of loci.

I find conceptual metaphors in general fascinating, so I have particular fun playing with the metaphors that Hiding Waters uses. In particular, rather than the conduit metaphor, Hiding Waters uses a "gathering-birds" metaphor to model the relationship between language and information.

Setting

Hiding Waters the language is named after Hiding Waters the place: a coastal river complex in a temperate rainforest on the cusp between an oceanic and subarctic climate. The area is populated by a species called hulhxrịnụ̀whrụ́l ("people to whom skins were given"). Hulhxrịnụ̀whrụ́l occupy a range much wider than just Hiding Waters. While human-like, they differ from humans in a number of ways which influence their languages:

• Hulhxrịnụ̀whrụ́l have the natural ability to transform between two shapes: one fairly humanoid, the other something else. The families of Hiding Waters are wolves in their second shape; in other regions, there are families of orcas, ravens, bears, and others.
• In both their shapes, hulhxrịnụ̀whrụ́l have highly-developed senses of smell. Where humans might speak of "mental imagery", hulhxrịnụ̀whrụ́l have a mental scent-scape.
• Hulhxrịnụ̀whrụ́l vocal tracts are mostly human-like in their humanoid shape. Similarly, in their second shape, their vocal tracts are not quite identical to their single-form counterparts. Their languages tend to have a second phonological register used in the second shape.

Stance

An omnipresent distinction present in the Hiding Waters language is the concept of stance. In any given interaction, Hiding Waters speakers have one of two stances with respect to each other: either leading or following. The two stances always contrast—when one speaker uses leading stance, the other uses following stance.

Stance marking in the language is pervasive. Each predicate and some function words inflect to mark the speaker's stance toward the audience, and predicate arguments inflect to show the speaker's stance toward their referent. For 1st-person references, or referents whose stance toward the speaker is unknown, the speaker's stance toward the audience is used instead.

Stance relationships are seldom static. While speakers may maintain the same stance relation over an entire conversation, they more typically switch back and forth as the conversation unfolds in response to various factors.

Leading Stance

Speakers who are relatively older or in a position of authority relative to their listeners will typically use leading stance. Similarly, speakers use leading stance when they have some form of territorial claim where the conversation is occurring, such as when hosting a visitor in their home, or speaking to a foreigner in their family's territory. Even having arrived first to a common area confers a weak territorial claim for purposes of establishing initial stance.

Leading stance may also indicate that the speaker has relatively more expertise in the topic of conversation. In situations where there is not a clear age, rank, or territorial disparity such that stance is more fluid, speakers may switch into leading stance as a way of requesting a conversational turn. (The listener then switches to following stance to cede the turn, or continues using leading to reject the turn switch.)

Following Stance

Speakers who are relatively younger, of lesser social rank, or in some sense "visiting" in the area where the conversation is occurring use following stance. Following stance may indicate that the speaker is petitioning the listener for their expertise, or, in circumstances where stance is more fluid (for example, two speakers of approximately equal age and rank speaking in a common area), speakers may switch from leading into following stance at the end of a statement to indicate that they are ceding the conversational turn to someone else.

Inanimate Stance

The inanimate stance is strictly linguistic; that is, culturally, speakers only recognize two stances, but there is a clear third paradigm in the language.

The inanimate stance is used to mark referents such as abstract or inanimate objects toward which the speaker has no stance. Plants, non-living objects, periods of time, and minor geographical features are all typically referred to using the inanimate stance. The inanimate is never used to mark the speaker's stance toward the audience.

Animals are typically referred to using an animate stance, along with some major geographical features and natural forces. However, some animals like mollusks and mosquitos are referred to as inanimate.

Phonology

This chapter gives a rough overview of Hiding Waters phonology, including its phonemic inventory, allophony, and morphophonological processes.

The language's sound system has a few notable traits that derive from features of hulhxrịnụ̀whrụ́l vocal tracts, such as a lack of any labial sounds, a lack of consonants which require grooving the tongue, and a lack of any phonemic rounding distinctions.

Consonants

Hiding Waters makes phonemic distinctions across three points of articulation. The inventory includes a series of ejectives and a fairly large number of sonorants, some of which feature a creaky-voice distinction.

Release

The manner in which consonants release varies depending on their environment.

When a plosive is followed by a fricative of like voicing at the same place of articulation (/tʃ/, /tɬ/, /t'ʃ/, /t'ɬ/, /dʒ/, /dɮ/, /kx/, /gɣ/, /k'x/, /g'ɣ/), the cluster is realized as an affricate.

• tulhlọtsị [tɯ˧ɬlɤ˩t͡ʃi˩] "it is small"
• ṭlhaùwh [t͜'ɬaɯ̯˥˩ɰ̊] "so many"

When any consonant is followed by an identical consonant not across a word boundary, the cluster is realized as a geminate.

• lulsisstẹ [lɯ˧lʃi˧ʃ:te˩] "she was talking"
• ngịghukkụx [ŋi˩ɣɯ˧k:ɯ˩x] "you got sick"

When a plosive is followed by a different plosive or fricative, the first does not release.

• sq̇ulkukqọq [ʃq'ɯ˧lkɯ˧k̚qɤ˩q̚] "he ate (the food)"
• sụruókjụk [ʃɯ˩rɯɤ̯˩˥k̚ɮɯ˩k̚] "I want to get strong"

When a plosive falls at the end of a word, it is not released, unless it is preceded by another plosive.

• sq̇ujukqọq [ʃq'ɯ˧ɮɯ˧k̚qɤ˩q̚] "I ate (the food)"
• sq̇ujukq [ʃq'ɯ˧ɮɯ˧k̚q˭] "I ate"

Vowels

Monophthong realization is conditioned by subsequent consonants:

• i > ɪ / _x, _q
• i > ɪ / _C[+creaky]
• e > ə / _x, _q
• a > ɑ / _q

Nasals and creaky-voiced sonorants assimilate with preceding vowels:

• V > V[+nasal] / _C[+nasal]
• V > V[+creaky] / _C[+creaky]

A number of diphthongs occur:

When two vowels are juxtaposed (noting that for these purposes, diphthongs are a single vowel), an epenthetic glottal stop is placed between them.

• xụlqu'ȯṇàlkulhq [xɯ˩lqɯ˧ʔɤ̰˥n̰a˥˩lkɯ˧lhq] "he is a good hunter"

Tones

Diphthongs always have a falling or rising tone. Monophthongs with a falling or rising tone are realized as long, and sometimes diphthongize:

• a[+non-flat tone] > a:
• u[+non-flat tone] > ɯʊ̯
• i[+non-flat tone] > iɪ̯
• e[+non-flat tone] > e:
• o[+non-flat tone] > oʌ̯

When a high-flat tone vowel occurs at the end of a word, it is followed by a glottal stop:

• V[+high-flat] > _ʔ / _#

Predicate Morphology

Predicates are by far the most morphologically complex part of speech in the Hiding Waters language. This chapter discusses the internal structure of predicates; later chapters expand on interactions and constructions involving multiple predicates.

Roots

Predicate roots are bipartite, consisting of a pre-stem and post-stem, although in some roots one of the two stems may be null. Most inflections are placed between the stems, although ablative and lative incorporated arguments are placed before the pre-stem and after the post-stem, respectively.

In the lexicon, bare roots are written with an asterisk (*) separating the two stems. For example:

ẹṇ*kwh

(atelic) a camp

In glosses, I notate the material between the stems as a series of infixes. For example, with the root ụng*q:

For consistency, I also use infix notation for roots with a null pre or post-stem. For example, with the root sqeị́s*:

Predicate Structure

The morphological structure of a predicate can be described with the following template, where elements in (parentheses) are non-obligatory. Subsequent sections will discuss each part.

(ABL) - pre_stem - STANCE.MOOD - (VIA) - (AGENT) - (NEG) - ASPECT - (PATIENT) - (LOC) - post_stem - (LAT)

Stance and Mood

The first morpheme after a predicate's pre-stem indicates the speaker's stance toward the audience—either leading or following—and one of six moods.

Indicative Mood

The indicative is the most common mood, used for general statements of fact.

Subjunctive Mood

The subjunctive marks a predicate that is not a statement of fact, but rather conditional, hypothetical, or otherwise counterfactual.

Optative Mood

The optative indicates something that the speaker hopes for or would like to happen.

Volitional Mood

The volitional indicates something that the participants marked on the predicate desire to do.

Potential Mood

The potential indicates something that is possible, that the participants are able to carry out.

Imperative Mood

The imperative is used for giving commands or exhortations.

Aspect

Predicates are inflected with one of seven aspects. For the most part, a predicate's aspect indicates its temporal structure—for example, whether the thing being described is a static condition, an ongoing process, or a completed change. Some aspects instead encode the speaker's interpretation of the situation, leaving its temporal structure up to context.

The seven aspects can be grouped into three families whose members serve similar functions and exhibit similar behavior in discourse structure: descriptive aspects, eventive aspects, and interpretive aspects.

(See also: Aspect Inflection Reference in the appendix)

Telicity

Predicate roots can be grouped into two families: atelic roots and telic roots. The telicity of a given root impacts the meaning of some aspects when applied to a that root.

Atelic roots denote some manner of quality which something may have or not have, and which may change, but which has no dynamic process inherent to it. Some atelic roots include:

gh*tsị̀

root quality: being beneath

ḳ*tlh

root quality: being a tree

n*t

root quality: being seen

Telic roots denote a process, which unless halted proceeds to some inherent completion state. Telic roots include:

h*s

searching (completion state: finding the object)

ṭ*st

repairing (completion state: restoring the object)

xụlq*q

hunting (completion state: felling the object)

Descriptive Aspects

The two descriptive aspects are the stative and essential.

The stative aspect (-x-) describes an incidental, non-inherent state of something.

For atelic roots, the stative aspect describes something having the root quality, though not as part of its inherent nature.

For telic roots, the stative describes something undergoing the root process, but does not place any saliency on progress being made toward the completion state.

The essential aspect (-lh-) describes an inherent nature of something. For atelic roots, it indicates that something inherently has the root quality; for telic roots, it indicates that undergoing the root process is part of a thing's inherent nature.

The essential can be used to indicate profession:

Or the position of stationary objects:

Eventive Aspects

The three eventive aspects are the imperfective, perfective, and climactic.

The imperfective aspect (-s-) describes an ongoing, dynamic process.

For atelic roots, the imperfective describes being in the process of taking on the root quality.

For telic roots, the imperfective describes making progress toward the completion state.

Note the difference in connotation between this and the stative example above; both refer to being in the process of going to a place, but the imperfective highlights the dynamic process, the progress being made, while the stative does not.

The perfective aspect (-k-) describes a change in state or completion of a process.

For atelic roots, the perfective describes a completed change of having taken on the root quality.

For telic roots, the perfective describes a finished achievement of the root's completion state.

The climactic aspect (-ṭė-) indicates a state-change like the perfective that occurs in a moment of intense interest. It might be surprising to the speaker, or a focal point of a narrative arc.

Interpretive Aspects

The two interpretive aspects are the inferential and the equative.

The inferential aspect (-r-) describes something the speaker suspects to be true given available signs.

The equative aspect (-ḷ-) is similar, but it does not highlight the speaker's belief. It makes no statement about whether something has the root quality; it says only that it has the appearance or form of something with that quality.

Agent and Patient Marking

Agent and patient morphemes consist of an onset, which indicates a classifier, and a coda, which indicates the stance of the speaker toward the referent. When the referent is the 1st person, or when the speaker's stance toward the referent is unclear, the speaker's stance toward the audience is used. (See the Catalog of Classifiers and Agent and Patient Codas Reference in the appendix.)

Here, the agent marker is -lku-, consisting of lk- (male classifier) and -u (leading stance), and the patient marker is -ktụ-, consisting of kt- (ungulate classifier) and -ụ (leading stance).

Predicates may lack a patient:

Lack an agent:

Or lack any participants:

Explicit participant markers may also be omitted when the participants are clear from context.

Morphological Negation

Predicates may take a negative-polarity inflection -n- just before the aspect inflection. For atelic roots, this modifies the meaning of the root to be "not having the root quality".

The negative polarity inflection converts telic roots into atelic roots meaning "not undergoing the root process".

Incorporation

Predicates may incorporate other predicate roots at four sites in the predicate complex.

Incorporated roots take an inflection specific to the incorporation site, which agrees with the speaker's stance toward the referent. The inflection is placed between the incorporated root's stems. (See the Incorporated Root Inflection Reference in the appendix.)

Incorporation is not recursive; that is, an incorporated root cannot contain additional incorporated roots.

Ablative Position

Ablative arguments are incorporated before the pre-stem of the root, indicating a source, origin, cause, reference point, or something moved away from.

Vialis Position

Vialis arguments are incorporated after the stance/mood inflection, indicating a route, method, duration, or something moved through.

Locative Position

Locative arguments are incorporated before the post-stem, indicating a location, focus, topic, boundary, moment in time, possessor, or something moved within.

Lative Position

Lative arguments are incorporated after the post-stem, indicating a destination, direction, goal, result, or something moved toward.

Reduplication

Roots may undergo reduplication by copying the post-stem twice, and prefixing the copies before the pre-stem. If the post-stem is nonsyllabic, epenthetic vowels may be inserted between the copies. Also something funny happens with tone, but I haven't figured out the rule yet.

Often, a reduplicated root indicates a repeated action:

Here, the root hạkt*q, "strike", reduplicates into qụquhạkt*q, "strike repeatedly."

Reduplicated predicates may also indicate a continuation of action:

General Syntax

Newsworthiness

Predicates are arranged in decreasing order of newsworthiness. A constituent's newsworthiness is determined by a number of factors:

• New topics (that is, topics when presented as a topic change) are very newsworthy.
• Surprising pieces of information are more newsworthy.
• Information the speaker wishes to highlight is more newsworthy.
• Rhemes are more newsworthy than themes (except when the topic changes).
• Given information is less newsworthy.

This is an excerpt from a conversation about a quọ́quxtsị̀lh festival the family was planning to host. In context, this statement introduced the salmon as the new topic of conversation, making them highly newsworthy. The fact that the salmon were caught by the speaker and listener was a focal point of interest; the fact that they were caught in the speaker and listener's weir, and the final clarifying detail of the weir's location, were less newsworthy.

Conjunctions

There are three conjunctions used to indicate relationships between different predicates. Two of these (su̇/sï and u̇k/ïk) inflect for the speaker's stance toward the listener. Conjunctions are placed before each conjunct in the coordinate structure.

The conjunction su̇/sï indicates a surprising or notable addition or confluence.

The conjunction tï indicates a mutual exclusion, either offering the listener a choice between options or indicating that only one of several possibilities is true.

The conjunction u̇k/ïk indicates a contrast or tension between two phrases.

Demonstratives

Demonstratives indicate a deictic reference to one of the arguments (either a marked participant, or one of the incorporated arguments) of the predicate. Demonstratives inflect for number (singular or plural) and the speaker's stance toward the referent. (See the Demonstrative Particle Inflection Reference in the appendix.)

When the marked predicate has multiple arguments, several of which have the same stance toward the speaker as indicated on the demonstrative, which argument is targeted by the demonstrative is ambiguous. Disambiguation can be accomplished via classifier extraction.

Demonstratives are not an obligatory component of the phrase; their inclusion typically signals the introduction of a new reference in the conversation. Once established, the reference is maintained using a classifier rather than repetition of the demonstrative.

• ṭ- indicates something within arm's reach of the speaker.
• k- indicates something within arm's reach of the listener.
• n- indicates something the speaker can see.
• ts- indicates something the speaker can hear but not see.
• ịng- indicates something the speaker can smell but not see.
• wh- indicates abstract, far away, or imperceptible to the speaker.
• h- indicates a classifier reassignment.

Quantifiers

Quantifiers are particles which precede a predicate and either indicate the degree or extent of the state/process the predicate describes, or the quantity of one of the predicate's arguments.

When a quantifier modifies a specific argument of a predicate which has several, the intended target of the quantifier may be ambiguous. See Classifier Extraction.

Discourse Markers

Discourse markers are particles which appear at the beginning of prosodic phrases, lending some kind of interpretive color to the phrase as a whole.

The most common of these is nà, which may act as a filler-word or tonal softener, or signal a change of topic or start of a new thought.

The majority of discourse markers are less neutral. Sì, for example, indicates a feeling of regret of misfortune. The following was said by a child who had been tasked with helping process a haul of meat from a successful hunt:

Hė and Nȧ

Whereas most discourse markers and particles precede their heads, hė and nȧ follow them. Both, like discourse markers, can apply interpretive color to an entire phrase, but hė sometimes behaves like a demonstrative or quantifier, marking an individual element of the predicate it follows.

Hė indicates a supposition of agreement, tagging a phrase as something the speaker believes is jointly known/understood information.

Hė may also act akin to a definite article, marking some element of its head predicate as a reference to common-ground information. (See Classifier Assignment.)

Nȧ, rather than supposing agreement, seeks it. It marks something that the speaker expects or hopes the listener will agree with or accommodate, but is softer in connotation than hė. It is frequently used as a tonal "softener", especially in requests and recommendations.

Predicate Serialization

Two predicates may be serialized to indicate that they both describe aspects of the same action or condition. Serialized predicates are occur adjacent to each other and are marked with the same aspect.

When a series of serialized predicates are used to describe steps in a sequence, they appear in chronological order:

Anaphora

Classifiers

Hiding Waters' classifiers form the framework by which references are managed in the language. Each classifier has two morphological forms: an onset, used in agent and patient marking, and a root form, used for free-standing predicates and incorporated arguments. (See the Catalog of Classifiers in the appendix.)

The classifier system reflects the way that hulhxrịnụ̀whrụ́l perceive the world around them: where similar systems in other languages might assign a classifier to an object based on its shape, form, or material, Hiding Waters assigns its classifiers based primarily on smell.

Each classifier is associated with a family of scent traces. Importantly, the scent traces on an object are not inherent to it, and a given object has a great many different scents on it at any one time. So, a person that has been working with fish may be tagged with the sl- "fish" classifier, or with the r- "person" classifier, or any of several others.

In the absence of other constraints, speakers pick which classifier to use for a particular referent based on which part of the referent's context they wish to highlight. Additionally, having so many classifiers to choose from provides a tool for disambiguation, as when introducing a new referent into a conversation, speakers will typically choose a classifier that is not already occupied by something else.

(These same choices are made when the referent is abstract, or not physically present. Whether or not speakers can literally perceive the smell traces on something, they can choose classifiers based on what they want to highlight, and what open classifiers are available.)

The language features a number of mechanisms to signal the assignment of a classifier to a newly-introduced referent, or the reassignment of a previously occupied classifier. Classifiers also provide a way to expand the otherwise constrained expressive possibilities of predicates.

Classifier Assignment

Speakers signal the introduction of a new referent with a new classifier via a number of different strategies, the foremost of which are predicate fronting, demonstrative marking, and the definite particle hė.

The introduction of new referents often coincides with the introduction of a new conversational topic, so predicates containing new classifier assignments often fall in fronted, more-newsworthy positions.

Very frequently, new classifier assignments are marked with demonstratives, which are dropped after assignment. Thus, when a speaker opts to include a demonstrative, it signals that they are either trying to disambiguate something, or assigning a new classifier.

In the above example, the demonstrative whó signals the introduction of a new reference in the following predicate. The stance inflection of the demonstrative, and the fact that the lk- "male" classifier has already been introduced, helps indicate that the sl- "fish" classifier is the new reference. (Sometimes additional disambiguation is necessary; see Classifier Extraction.)

The definite particle hė can also be used to assign classifiers to new referents. In the most common such pattern, a predicate which uses a previously unassigned classifier but contains insufficient information to identify its referent is marked with hė. Then, a less newsworthy predicate follows, again referencing the new classifier—but this time with identifying information.

In the above example, the "female" classifier ls- is used first in the fronted predicate, but as the information is novel, it does not give the audience enough to identify the referent. The hė particle affirms that the referent is something the audience is aware of, and the following predicate provides enough information to clarify who is being talked about.

Classifier Reassignment

It is often useful to repurpose a given classifier to refer to something other than what it was first assigned to in the conversation. Most commonly, this happens organically—if it has been some time since the classifier has been used, or it is otherwise clear from context that the classifier is now being used to target a new referent, then one of the assignment structures detailed above can be employed as usual and the new assignment is understood.

However, in cases where it would not otherwise be clear that the classifier is not referring to the old referent, the h- demonstrative can be used to explicitly indicate the reassignment.

Classifier Extraction

The indivisibility of a predicate's internal structure can result in ambiguity regarding which part of a predicate is targeted by external marking. For example, the following could be interpreted a number of ways:

It may be clear from context which interpretation is intended. If not, classifier extraction provides a mechanism for disambiguation.

Classifier extraction entails the addition of a new predicate, usually the auxiliary predicate s*, which is serialized with the main predicate. Some element of the main predicate is then copied into the auxiliary, and the ambiguous markings are applied to the auxiliary instead. The marked auxiliary may go before or after the main predicate, depending on which is more newsworthy.

Thus, the two possible interpretations of the above sentence could be unambiguously expressed as follows:

Anaphoric Incorporation

There are a number of anaphoric roots which refer in some way to the overall action or state of another predicate. These can be used as an incorporated argument in a number of structures to connect predicates together into more complex statements.

The most common anaphoric root is s*, which is also the most semantically blank. Other anaphoric roots are more specialized:

The root s*wh refers to the manner of the referenced predicate's action or state. In the example below, s*wh is incorporated as a vialis argument in susu̇whlkux, "he was doing it in that way".

The root s*kwh refers to the location of the referenced predicate. In the example below, s*kwh is incorporated as a locative argument in lixlọsokwhkwh, "it is in that place".

The root s*tlh refers to the time of the referenced predicate. In the example below, s*tlh is incorporated as a locative argument on qȧṭulkuklọsutlhkoụ̀n, "he stole it at that time."

Stance Agreement

The inflection on incorporated anaphoric roots agrees with the stance of the referenced predicate in the following ways:

• If the referenced predicate has neither a marked agent nor patient, the anaphoric root is inanimate.
• If the referenced predicate has one marked participant, then the anaphoric root agrees with the stance of that participant.
• If the referenced predicate has both a marked agent and patient, then the anaphoric root agrees with the stance of the more salient participant.

Causation

On predicates in the indicative mood, s* can be incorporated as a lative argument to mark a cause, or as an ablative argument to mark an effect. The causal predicate may come before or after the effected predicate, depending on which is more newsworthy.

In the above example, the auxiliary refers to the predicate ḳuṇunks, and is incorporated as a lative argument on the predicate awhuxsn to mark it as the cause of the thing the auxiliary refers to.

In this example, the auxiliary refers to the predicate lijikstẹhị́lịk, and this time is incorporated as an ablative argument on xàrixlkịjing to mark it as the effect of the referred cause.

Conditional Structures

The same causative structure can be used with predicates in the subjunctive mood to form conditional statements, with s* incorporated as a lative argument to mark a protasis, and as an ablative argument to mark an apodosis. (Only one of the protasis or apodosis is typically marked.)

Indirect Speech

Indirect speech may be referenced by s*wh, incorporated as a vialis argument on l*shte, "saying".

Specialized Structures

This chapter contains descriptions of an assortment of interesting structures the language uses for various purposes.

Affirmation and Negation

The particles saù/saì and u̇ng/ïng are used respectively to affirm and negate predicates.

The u̇ng/ïng particle differs from the -n- negative infix in that it negates the entire inflected predicate, rather than just the underlying root. For example, compare the following:

Questions

Questions in Hiding Waters are formed via a collection of particles. Question-phrases do not exhibit any unique behavior in terms of word order; they follow the usual newsworthiness rule with respect to ordering of predicates (although, since the focus of a question is often highly salient or the subject of a topic change, questioned elements frequently occur in fronted positions).

Question particles almost always follow the predicate they target. Since question particles may target a specific element of the predicate rather than the predicate as a whole, they are often used with classifier extraction for disambiguation.

Polar Questions

Polar questions—that is, questions whose answer is either "yes" or "no"—are formed with the particle tí.

Content Questions

Content or "wh"-questions are formed with the particle hí.

Note the absence of a specialized question word like "when". Hí often coincides with some manner of deictic reference in the predicate, such as n*tlh "recent/soon" in this example, which serves as the target of the question.

Alternative Questions

The question particle hí is used in combination with the conjunction tï to ask the listener to choose between a set of options.

Notice how the predicate lusu̇whkstẹkụ́ omits the usual agent marking; this signals the "blank" which the questioned elements are intended to fill.

Note also the usage of tï in the response. This is not obligatory when responding to an alternative question. It is used here only for emphasis, to reinforce that only one of the two told the speaker.

Leading Questions

The affirmative particle saù/saì and the negative particle u̇ng/ïng can be used with the polar question particle tí to form leading questions.

Topic-Shifting Questions

The question particle ha is used to reapply a recent question to some new target.

Selected glosses:

Copulative Phrases

A copulative phrase is used to highlight that two references both point to one-and-the-same referent. It is one of few cases in which the particle hė is placed after another particle, rather than a predicate.

The structure consists of a predicate phrase which contains a reference that has already been established by more newsworthy predicates earlier in the phrase. This less-newsworhty predicate phrase is preceded by tï hė, which highlights that the referent it points to "...is the very same one that..."

Comparative Structures

A number of different comparative structures may be formed using incorporated anaphoric roots and specialized comparative roots.

More and less

The comparative roots *ṇ and *tsị̀x are used to indicate that something is more or less than some basis of comparison, respectively. The basis of comparison, or an anaphoric reference to the basis of comparison, is incorporated as an ablative argument, and the appropriate comparative root is incorporated as a lative argument.

In situations where the ablative or lative position is already occupied, classifier extraction can be used to form the comparative structure.

Not as

A predicate marked with the negative particle u̇ng or ïng may incorporate a root referring to a basis of comparison as a lative argument to indicate that it is to a lesser extent than the basis.

Equally

One way to say that something is "as X as" something else is to incorporate the basis of comparison as a locative argument.

Another strategy is to serialize the predicate describing the common characteristic with the root t*sx, and incorporate the basis for comparison as a locative argument on t*sx.

Experiential Structures

In experiential statements, the experiencer is marked as a locative argument on a root indicating the form of the experience; the subject of the experience is marked as a patient.

To describe an experience of something in a location, the location is marked as an ablative argument rather than locative.

Discourse Structure

This chapter details some interesting dynamics in the structure of Hiding Waters discourse; in particular, it examines how stance-switching is used to signal conversational turning points, and how the distribution of aspects tracks the trajectory of narratives.

Stance Dynamics in Discourse

As discussed in the chapter on stance, in some situations stance behaves like an asymmetric formality register, with one speaker holding leading stance while the other holds following stance for the duration of their speech. However, much more interesting dynamics emerge when the scales are more evenly balanced—when there is no clear territorial claim or difference in rank holding the speakers' stances in a single configuration.

The following is a dialogue between two speakers of similar age talking in an outdoor common work area. Speech in leading stance and in following stance is marked respectively.

Line by line glosses:

This next text is a dialogue between two young men who are both newcomers to the clan.

Line by line glosses:

The root qilh*q̇ẹ̀ specifically refers to a bounding, loping gallop, as of a wolf or hare. Qilh…sqọ'osnq̇ẹ̀, then, means "making stones bound on water".

The above sentence includes an interesting structure: hï tï huksqọtlh, "which (stone) is chosen".

The particle hï precedes a phrase to indicate an embedded question, such as Huxjukwh hï inguxlọkwh, "I know where it is", or Hï sijis laìkikstẹjị́, "Tell me what to do." This usage of hï occurs as well with the root h*tlh, "choose", without the inclusion of tï. For example: Hï hoàjuktlh hunxjukwh, "I don't know what I should pick."

The inclusion of tï above serves to highlight the importance of the particular choice, at the exclusion of other options.

The above sentence uses a derivational strategy for describing things with multiple traits: lịxuslhȯwhịtlhsqọsịq, "(the stone) is flat and smooth", is formed by incorporating the root for "smooth" as a vialis argument on the root for "flat", creating a predicate meaning both traits together.

Conceptual Metaphors

Gathering Birds

There is a widespread metaphor in Hiding Waters concerning language and thought, which surfaces in a broad range of core expressions:

The root of huhnȯwhkulhsxọtsị̀lh ("with those imperceptible things, you gather them") is used to describe an action one does that causes something to gather in a particular place, such as building a weir to catch fish, or playing music to gather a crowd, or attracting animals by leaving out food.

The root of jị̀kịtsiwhoksxọḷȧ ("Those imperceptible things scattered them away from me") describes causing a group of people or animals to scatter away from a place, such as scaring fish away from a boat, or making a sound that causes a flock of birds to take flight.

The root of ịtọ̀sxḷeìxjikwh ("I can infer from those bird-smelling things") refers to inferring something from an available sign or trace, such as making an inference from the shape of a footprint, or inferring the proximity of a predator from a change in bird calls.

It is notable that throughout these examples, and indeed in general, the preferred scent-classifier for referring to language signals such as words, songs, or other speech is the same classifier used to refer to scent-traces of birds.

These examples suggest a conceptual metaphor in Hiding Waters for the relationship between language and information:

Certain changes in the environment cause certain birds to make certain calls. A trained listener, upon hearing these calls, can infer the presence of the associated environmental change. Language signals are like birds: a trained listener may, upon hearing them, make accurate inferences about otherwise imperceptible facts.

An interesting quirk of the metaphor is that language signals are not spoken of like bird calls, but rather like the birds themselves. The speaker does not themselves make the birdcalls—rather, the speaker gathers the birds, and the calls follow on their own.

As described above, "gather" is a tricky translation here. The English translation overlaps with the kind of "gathering" one does with berries or firewood, whereas h*tsị̀lh does not. Instead, it refers to the sort of gathering whereby one shapes an environment such that something will collect there, such as digging a cistern to gather rainwater, or smoking a catch of fish to gather one's friends.

Where the speaker is conceptualized as gathering the words, the listener is conceptualized as interpreting the presence and nature of the words to understand something about where and to what they have gathered.

Those inferences, the information which the listener comes to understand, may be spoken of as locations where language signals gather, or as bait which attracts them.

Time is a Landscape

Another widespread metaphor conceptualizes time as a landscape through which speakers move. This metaphor manifests in core expressions like the following:

The root qẹ̀ṇ*xtẹ prototypically refers to venturing off of the usual path in other to take a shorter route.

The root xọ́q*ngạx, incorporated here as a vialis argument, prototypically refers to a winding route, a great travel distance.

The root k*sụx refers to the near side of some reference point, which is incorporated as an ablative argument. K*sụx may be used either geographically, indicating that something is between the reference point and the speaker's current position, or temporally, indicating that an event either occurred after the referenced event in the past, or will occur before the referenced event in the future.

Appendix

List of Abbreviations

Inflection Reference

Stance and Mood

Agent and Patient Codas

Aspect

Incorporated Root Inflections

Demonstrative Particle Inflections

Catalog of Classifiers

Catalog of Discourse Markers

hà

Indicating understanding/knowledge of something

hė

Post-predicate marker, supposing agreement, indicating jointly-held information

hoù

Indicating concern or awe

nà

Phrasal marker, continuation, softener, request for conversational turn

nȧ

Post-predicate marker, softener, diminutive, seeking agreement

sì

Indicating something unfortunate, regretful

uá

Indicating surprise

ué

Calling attention

xaù

Indicating disgust

Catalog of Quantifiers

hï

any, whichever

kxí

some, a few

ḷȧwh

whole, complete

ṇė

all, every

ngín

mostly, for the most part

ȯx

a moderate amount, fairly

qė

always

qïx

just, merely, a surprising scarcity

síx

a little, barely

tï

only, one

xu̇

much, very

àṇ

more

ïlh

some, only partly

ṭlhaùwh

so much, so many, a surprisingly large amount

Lexicon

*ṇ

(atelic) more

*ṇà

(atelic) good

awh*sn

(atelic) rain

axṭ*ḷị́wh

(atelic) tall

ạng*hȧ

(telic) sing

ạṭ*nịq

(atelic) sitting

dịj*lhnịn

(atelic) the halfway point, as of a journey or crossing

dị́l*xtẹ

(atelic) far off, a long way to go

etlh*koùn

(telic) take, receive, accept

ẹṇ*kwh

(atelic) a camp

ẹṇ*sịlh

(atelic) inhabit, live in a place

g*wh

(atelic) decent, acceptable

gh*ng

(atelic) big

gh*tsị̀

(atelic) beneath

ghị́gh*x

(atelic) high quality, as of food

h*koụ̀n

(telic) carry

h*kwh

(atelic) known

h*lịs

(atelic) female classifier

h*ḷì

(telic) offer, present

h*naụ̀wh

(telic) cousin, sibling, family member of like generation

h*nụ̀wh

(telic) formally give

h*q̇ė

(telic) great exertion

h*s

(telic) searching

h*tlh

(telic) choose

h*tsị̀lh

(telic) gather, collect, attract

h*xtẹ

(telic) scout, explore

hen*x

(atelic) everyone

hạkt*q

(telic) strike

hịng*t

(telic) lure, distract, bait

hụkt*dlụ̀wh

(atelic) a fishing weir

hụkt*q

(telic) catching, capturing

hïn*stẹ

(atelic) welcome

ing*wh

(atelic) healthy, well-being

inx*whn

(atelic) upriver

ịng*xtẹ

(atelic) away, elsewhere

ịt*kwh

(atelic) beside

ịwh*xk

(atelic) knife

k*koị̀

(atelic) father

k*lẹ

(telic) continuing, persisting until completed

k*lịt

(atelic) yesterday

k*lịx

(atelic) far side of

k*qụ

(telic) finished

k*q̇ė

(atelic) extreme, a great extent

k*sụx

(atelic) near side of

k*xtẹ

(telic) going

kịts*lȧ

(telic) scatter

kx*ghẹ̀

(atelic) negative extreme

kx*ts

(atelic) damaged

ḳ*doq

(atelic) last night

ḳ*s

(telic) work

ḳ*tlh

(atelic) tree

l*kwh

(atelic) place, being in a particular place

l*ste

(telic) spoken

l*sq

(telic) taught

lịx*sịq

(atelic) flat

lh*sn

(telic) stream, small river

lhị́s*xẹn

(atelic) encircled

lhị́sq*ṇguq

(atelic) curved

lhq*tsìwh

(atelic) cedar

ḷ*kwh

(telic) infer from a sign or trace

ḷàng*stẹ

(telic) formal speech

n*

(atelic) two

n*s

(telic) help

n*t

(atelic) seen

n*kwh

(atelic) place within view

n*ṇoị̀

(atelic) mother

n*tlh

(atelic) near time, recent or soon

ng*sịlh

(atelic) here; region where the speaker currently is

ṇ*sịlh

(atelic) home region

nạ́gh*xn

(atelic) salmon

ng*l

(atelic) much, a lot

ng*sịlh

(atelic) here, general area around the speaker

ṇg*zẹ̀

(atelic) feeling emotion

ngịgh*x

(atelic) sick, ill

or*koụ̀n

(atelic) hold, possess

qaqax*stẹ

(atelic) chattering, squawking

qȧṭ*koụ̀n

(telic) steal

qạx*l

(atelic) berry

qẹ̀ṇ*xtẹ

(telic) shortcut, take a shorter route than the usual one

qẹ́ts*s

(atelic) neck

qilh*q̇ẹ̀

(atelic) running, galloping

quọ́q*tsị̀lh

(atelic) giving feast

q̇*ṇgeị̀

(atelic) nourish

r*t

(atelic) self, alone

s*

(telic) semantically blank auxiliary; "doing"

s*wh

(atelic) anaphoric root referring to the way or manner described in some other predicate; "thusly"

slh*whịt

(atelic) smooth

slh*x

(atelic) mush, paste

stsịn*qt

(telic) urinate

sṭȧn*ti

(telic) dance (event)

sṭlh*

(telic) try, endeavor to

sụr*k

(atelic) strong

sqeị́s*

(telic) beginning

sq̇*q

(atelic) eat

sx*nịq

(atelic) white

ṭȧn*ti

(atelic) dancing

tạṇ*tlh

(atelic) a house

tsạk*xtẹ

(atelic) traveling

t*

(atelic) one

t*ngạk

(atelic) finger

t*ṭạk

(telic) jump

t*qọ

(telic) throw

t*sx

(atelic) same

t*tsi

(atelic) small

tị̀n*rọq

(atelic) glaring, glowering

tlhị́lh*skẹ

(atelic) clean

ts*wh

(atelic) heard

tsụn*hị́lh

(atelic) swaying

ṭ*doq

(atelic) tonight

ṭ*kwh

(atelic) here, space immediately around the speaker

ṭ*lịt

(atelic) today

ṭ*st

(telic) repaired

ṭ*tlh

(atelic) now

ṭeị̀n*tsịq

(atelic) kinked, crooked, as of a rope, bough, or joint

ṭsẹ́s*slh

(atelic) a woven mat

ṭị́ng*tsi

(atelic) beside

ụng*q

(atelic) hungry, envious

wh*kwh

(atelic) location outside of scent range

whịn*t

(atelic) light, not heavy

xàr*ng

(atelic) angry

xing*kwh

(atelic) thinking

xix*jn

(atelic) spoiled, gone bad (as in food)

xịlht*q

(telic) bite

xng*sq

(telic) smoke meat

xoụ̀n*sqọq

(atelic) jacket

xọ́q*ngạx

(atelic) a winding route, a long way to travel

xur*kw

(atelic) south

xụlq*koụ̀n

(telic) acquire

xụlq*q

(telic) hunt