In this post, I stumble through design decisions for a constructed language with no syntax and too damn much morphology. I'll post a compact reference grammar around April when it's done.
How about some phonology? The Xenants don't have vocal cords, and neither have they vowels or voiced consonants. Awesome, right? A Xenant has one mouth-like organ on each side of its face and can produce sounds from both mouths simultaneously, independently. If you're not excited yet, just walk away. The mouth-like organs are only used for making noises. The orifice for eating is instead on a Xenant's thorax. Or maybe I'll switch the locations. A cicada's noise producing organs (tymbals) are located on its abdomens after all, and likewise with the katydid's tegmina. Whatever. Either way.
What kind of consonant inventory would be right for a crystalline insectoid alien? Different anatomy means we could have all sorts of new sounds. Imagine a bird's trills, but made by an angle grinder. But I won't do that. It would be awesome in a theoretical sense, but practically, it might interfere with appreciating all the other things I want to put into this. Like, I considered that they could have mouthparts of stone, metal, and glass, with novel sounds arising from using those against each other. But I wouldn't begin to be able to read that or parse it if I heard it. Tonight, I think the Xenants should have sounds like our Ch, T, and K, i.e. the voiceless postalveolar affricate ⟨tʃ⟩, the voiceless alveolar plosive ⟨t⟩, and the voiceless velar plosive ⟨k⟩. Clicks are kind of alien and also make sense for crystalline percussive mouths. I don't know that I want them though. Let me sleep on it.
They can also make a scraping noise, a bit like our Sh, the voiceless palato-alveolar fricative ⟨ʃ⟩, but they only use it as a non-linguistic speech sound.
I want all of the consonants to have two forms, high and low. If you make a Ch noise with your lips and tongue in the shape for ⟨i⟩, the close front unrounded vowel, it will sound like the Xenant high-Ch, and if you make a Ch noise with your lips and tongue in the shape of ⟨u⟩, the close back rounded vowel, that will sound like the Xenant low-Ch. I don't know what the right word for that in phonology. I'll call it tension. Or height. It might just be pitch? Stops and affricates don't have easily recognized fundamental frequencies, and that's what we normally call pitch, but you can still move the sounds up and down the frequency spectrum, giving high and low variants. Sorry if there's a standard term I'm not using. I don't know much phonology; languages are cool to me because they express thoughts, not because I like thinking about flexing my mouth meat. You can do an even lower-height consonant by putting your mouth in the form for a near-open central vowel ⟨ɐ⟩ or an open-mid back unrounded vowel ⟨ʌ⟩, so we could have high, medium, and low height consonants, but let us not do that. Actually, I tried with a bunch of different vowels, and I think I consistently perceive an order like ki > ke > ku > ko > ka.
I'm told by the fine conlanging people on discord that the high version of Xenant consonants might be palatalized (as happens allophonically in English before front vowels and the palatal approximant), and the low version might be rounded (rather than labialized, even though roundedness usually describes vowels), and the very low version which I'm not using might be plain, as in non-palatal and unrounded. They also suggested that it would be natural for me to distinguish velarized consonants from palatalized consonants, as Irish does, for better distinguishability. Very interesting, but I probably won't.
The high and low forms do not have any semantic relation to Xenants, any more than voiced and unvoiced consonants are related in English across the words "god" and "cot".
So we've got (3 * 2 = ) six consonants now, but Xenants can make two at a time, and maybe the pair of sounds is what we should consider a consonant. But if K is a consonant, and Low-K is a consonant, and (Low-K, High-Ch) is a consonant, we're seriously overloading that word. So how about this: Ch, T, and K, with height unspecified, are "articulants". When you specify the height, you've got a consonant, and when you've got two of them at once, you've got a syllable. Without vowels in the language, nothing else is going to be considered a syllable, so we might as well put that word to good use.
Here's a compact ASCII coding: Upper case is for the high consonants, lower case is for the low, T and K are written as such, but Ch is written with an X or an x.
One more symbol I'd like to introduce is for silence: it'll take some energy for either mouth to make a noise, so it makes sense as a principle of energetic economy that the Xenants would often make one sound instead of two. I'll transcribe a silence with a period. I won't use silence in both mouths at once as a syllable. Our maximal syllable structure is now C(C). I think that's pretty funny.
Here's a Xenant word in three syllables, with dashes to separate: "Kx-xk-K." For the humans among us, we could write that more pronounceably as kichu-chuku-ki', and we'll just have to learn to accommodate their halting one-mouthed speech impediment. If you haven't done, try saying it without any voicing. It's kind of fun.
I've written Kx as the first syllable. You might wonder if K is spoken in the left mouth or the right. I've decided that whether a consonant is produced in the left mouth or the right will not influence the meaning of a word, because I imagine the difference would be hard for listeners to disambiguate, even for alien listeners. Thus the language doesn't specify where the K is made. Ants with a dominant left side might read the first consonant of every word with their left mouth and the second with their right. An ambidextrous ant might alternate, between each word, which mouth gets the first written consonant, so as not to overuse and tire either mouth. Oh gosh, maybe the Xenants don't have anything like cerebral lateralization. Why do Earthicans even have it? Hmph.
How about some phonotactics?
It's a little hard for me to say "t-t-t-t" rapidly, while I can say "ch-k-t-k" like the wind. So let's suppose that Xenant words generally avoid putting the same articulant in one mouth on successive syllables of a word. That's more pleasing to the ear and to the tongue, and it's easier for the listener to determine what's an intentional utterance versus a hesitation or a stutter. It's also not very human. Infants prefer words with repeated syllables ("dada", "moo moo") and reduplication is a productive tool for deriving new words in lots of languages. But not mine.
Honestly, after a little practice I can also say "t-t-t-t" pretty quickly, but I need some phonotactic constraints, and this one has some justification.
That's it for traditional, horizontal, sequential temporal phonotactics. I also have some vertical, spectral, instantaneous constraints:
It's probably easier to make the same sound with both mouths than to make two different sounds, so the core vocabulary should make a little more use of TT as a syllable than Tk for example. A half silent syllable "T." is still favored more than either of those.
Secondly, I like the idea that a syllable shouldn't have the same articulant in both mouths while having different heights (so TT and tt are fine, but not Tt). I don't have an explanation for that restriction in terms of how it would help with communication, so I might weaken the rule at some point. For now, I'm not considering any words that break it.
Ooh, or the core vocabulary of semantic roots could avoid syllables of mixed height on a single articulant, and then I could use them for particles or relative pronouns or something.
I've already transcribed Xenant words two ways (Kx-xk-K. and kichu-chuku-ki'), but how about a third scheme to finish this little introduction off? If we use the following key to make substitutions ("K": "┐", "k": "┘", "T": "┌", "t": "└", "X": "─", "x": "│", ".": "•", "-": "",) of letters for box-drawing characters (and a bullet point for silence), then we get a beautiful script:
─┌┘│└•┘•── ┐•┌┌┐•│•┘┌──┘• ┌•─•┘•─•└•┘• ┐•─•┘•┌•─•┌•┐┌ │┌┘│┌•│•└•┐│┌• ┘•─└┐•│•└•┐• └┐┘─└•┘┌│•└•│• ││┘┘│•┌┘─└ └•┘─└┐ ─•┘┘└•┘•│•
Does it mean anything? No. Is it even easy to see the word breaks? Not at all. Is it alien and awesome? Hell yes. It looks even better in my text editor; the strokes join together across the lines of text, forming a beautiful 2d maze. I'd argue that it's also a featural script, which is nice.
I'll make a new script some time that's easier to read while being no less pretty. Perhaps one where the simultaneous consonants of a syllable are joined up into one character. I suppose without vowels that would be a syllabary and an abugida at once. You might have seen the 3d box script that I posted on my @postinfarction twitter account last May. I might adapt and adopt that for the present purpose.
That's a good start for tonight. Goodnight.
-
Day two. I've decided that bilabial clicks (kissing noises) are dental clicks (the tutting, tsking noises) are cool and ant-like, while alveolar, lateral, palatal clicks are cool but not ant-like. If you think bilabial clicks are not ant-like, I encourage you to listen to some recordings of actual ant chirps (stridulations).
A bull's eye is often used for transcribing bilabial clicks, so I'll use the similar O and o in the compact ASCII coding. The lower-case o is a familiar motion, like blowing someone a kiss, while the upper-case high consonant is weirder and is basically done by breaking a smile. Sometimes you hear people make it before saying, "well" or "welp" as an interjection. In the more coding scheme used to guide human pronunciation (the romanization, I suppose), O and o will be written pi and pu.
A vertical bar is often used to transcribe dental clicks. I'll use I and i to approximate that for the compact ASCII coding. In the pronounceable coding, they'll be written zi and zu, I guess. I had been doing "pchi" and "pchu" and "tsi" and "tsu" for the stops, which sound closer to the ant sounds, but whatever, short names are fine.
What are the phonotactics for the clicks? The clicks are all unfamiliar to my mouth, and I have a hard time putting them in any sequence with other sounds. I'm going to practice a little and see which sounds I *can* with practice put in sequence before saying which sounds shouldn't be put in sequence because they're too hard for me to do, and thus perhaps too hard the aliens. I think the bilabial click and the voiceless postalveolar affricate (Ch) probably won't occur in direct sequence in a mouth.
That's enough phonology for a little while. How about syntax? I'd like the Xenants to have a very free word order, long sentences, and deep nesting of clauses, as if they had larger working memory capacities than humans. I might balance that against some other mental limitation later, like a limited episodic memory. I guess I'll mostly address primary word order at first (the order of verbs and their arguments, like SOV).
Marking nouns for case and marking verbs for polypersonal agreement are two natural mechanisms that allow for freer word orders in human languages. I don't know enough about them in practice, but I think they probably still struggle to disambiguate the parse tree of a sentence with nested subclauses and similar nouns like "The dog who chases the chicken likes the cat who hates any dog who chases a chicken" after the order of the sentence has been forgotten. Or maybe not? I've seen some very long translated Latin sentences from Cicero, and he does not shy away from strange word orders or nested clauses, so maybe I should just learn more Latin or another syntactically liberated language. Or maybe the sentences really are ambiguous, and situational context is a significant part of their interpretation.
There's an easy enough way to make nested sentences unambiguous that doesn't require me to learn new things: we could introduce a unique marker for each noun, and have the verbs incorporate the markers for each noun that they operate on. Here's a marked version of the dog sentence, using numbers as markers: (1-the-dog 1,2-chases 2-the-chicken 1,3-likes 3-the-cat 3,4-hates 4-any-dog 4,5-chases 5-a-chicken). Put that in any order, and I can still tell you that the cat hates "any-dog", and that that same "any-dog" chases a chicken. And you could put numbers on the articles and quantifiers too, instead of glomming them onto their nouns, if you want to scatter those around for some reason.
It's ugly, I admit. It's something that you could read if you had a larger working memory, but is it something you'd want to compose? Why would the Xenants have a system such as this, other than that I want something new and weird in my life? Does it have some communicative benefit? Free word orders let you decide when you want to bring parts of a sentence to your audience's attention, of course. Is that it? I want a language that's weird because it's suited for someone different from me, even if I have to invent a new person for whom cool language features are suited. What communicative purpose or mental affordance could this suit, beyond letting someone mention "cat" at the beginning of the sentence and cat's definite article "the" at the end of the sentence?
Okay, that's not fair, even Latin uses free word order more productively than that. But I'm not sure what I want it to achieve for the Xenants.
Let me sleep on it. I really like reading translations of Cicero, even when he doesn't have anything good to say. It's just beautiful how he wends and wanders. I'll think of some justification for free word orders, and then other parts of the language will flow from that. Goodnight.
-
Day three. You know what other mechanism allows for a very free word order? A stack based grammar, as found in the programming language Forth and the conlang Fith. I would not seek to do Fith an injustice by making a pale imitation. But I might *accidentally* do it an injustice by making a pale imitation. And why should Fith be the only conlang with stacks? There should be many. It's a good mechanism.
You may know that a push-down automaton with two stacks is equivalent to a Turing machine. Perhaps I could make a language with two stacks. Two interacting mouths and two interacting stacks. Does your language not have a Turing machine in it? That's cute. I bet it's easy for children and pets to use.
Of course there will be some finite stack depth for a finite brain, and so this language, like any physically instantiated language, could technically be analyzed as being regular, and not recursively enumerable. This is known as "The Useless Theorem". I'm still doing two stacks though. Probably. My opinion changed in less than eight hours. We'll see what the next eight hours bring.
There are five kinds of phrasal disorganization (hyperbaton) enabled by free word in old Latin that were analyzed by the Roman grammarian Donatus: rhetorical parenthesis, synchysis, hysterologia, anaphora, and tmesis. I really like parenthetical remarks (of which here is a nice typology), and phrasal interleaving (synchysis). Hysterologia is alright; it means presenting events out of causal order in order to focus on effects. It's a little jarring as usually translated, but if you mark events as causes or effects, like with a preposition "by", then you can keep your intended topical focus without making the reader correct for the fact you don't know which way clocks spin. A fourth kind of hyperbaton spoken of by Donatus is anastrophe, which mostly consists of rearranging subjects, verbs, and objects. Anastrophe is associated with bad poetry in English, and with Yoda, but it doesn't sound out of place in a language that doesn't have a default word order. A little anastrophe is fine by me and useful for the Xenants. Finally tmesis means infixing words or phrases between morphemes of a word, which I don't want in my language, and it wouldn't naturally appear in a language simply as a consequence of having a free word order. Tmesis interleaves phrasal disorganization with lexical disorganization - which is a whole nother thing.
So, I will devise a two-stack grammar that supports writing cool sentences with rhetorical parenthesis and synchysis, and anaphora and synchysis are fine too. Nice. And if that doesn't work out, we can still doa one-stack grammar like Fith.
What about the lexicon? I'm not really ready for that. They'll have words for things in their world which suggests we should do some world-building. I'm not hugely into world-building. But let's try anyway. The Xenant homeworld does not have liquid water or carbon-based life. I should probably talk about what quirks of stellar nucleosynthesis and planetary formation caused that, but not everything needs an explanation. I can just say there's no water if I want to. The hydrogen got stripped off of their atmosphere by ionizing stellar radiation. Maybe the planet is highly radioactive and volcanic, and the Xenants evolved at the interface of the planet's core and lower mantle. And they have a stony or metallic carapace, but their internals are more like silicone polymers with alkali radical groups. I don't know. I shouldn't have to be an expert in speculative xenobiology to make a toy language.
How about biodiversity on the Xenant world? If there are intelligent motile organisms, there will probably also be less intelligent motile organisms and unintelligent sessile organisms.
I could look at words related to organisms on Earth and decide if there are close versions on the Xenant world. I don't know how to decide though. Should there be swimming/flying things? Should there be grazing animals, scavenging animals, and predatory animals? Ants are a given. Are there bees? Probably not if the atmosphere has been stripped away. Unless it's a magmatic bee, I guess. Which is more like a fish, really. Fish are just water bees, as the saying goes. Except now they'll be in magma. Hardly different. Water is just ice-magma. Did the alien ants co-evolve with alien plants as their pollinators, or do the ant eat the majority of the plants like herbivores? Or do they get energy from the plants in a different way besides ingestion?
And what about those plants? How am I supposed to decide what plant forms have a homologue in the alien ant language? Do their plants have anatomical parts like (berry bloom branch flower fruit leaf nut root stem)? Are their plants even photosynthetic, or are they perhaps chemotrophic or radiotrophic? How would that influence their anatomy? I can ask these questions, but I don't feel suited to answering them.
Oh, if Xenants evolved in a magmatic ecosystem, then we can say they probably don't have vision and their plants aren't photosynthetic. And if they're blind, then maybe their written script has some depth like braille or cuneiform. That's good. I like when new features follow as a consequence of things I've already settled on. And maybe they have an inhuman sense organ adapted to their environment as compensation. And maybe they're significantly less dense than the magma they inhabit, so they float up the underside of the planet's matnle (the core-mantle boundary), and their ecosystem seems upside down compared to ours.
Oh no, I just remembered that Terran ants will poop and vomit in each other's mouths as part of their social feeding, and that they communicate some information that way, particularly by the transfer of pheromones. Y'all know about trophallaxis? I don't know if earth ants have control over which messages they send that way, but maybe I should include something similar for the Xenants. You know, to compensate for their lack of sight. I could redesign them to eat vomit to make up for blinding them. That's the kind of benevolent creator I am.
Ooh, or the xenants could have magnetoreception and/or electroreception. Apparently the antennae of honey bees can function like an electroscope. Pretty cool. The Xenants could be covered in little hairs of asbestos and ferrite, maybe.
I still don't know what the animals or plants really look like, but that's okay. I didn't want to talk about fake animals and plants in my conlang anyway. I'd rather talk about things like (this big list of mostly abstract nouns you should feel free to skip):
* Ability, mastery, technique, competition, and strategy
* Agreement, common knowledge, deception, and secrecy
* Abundance, scarcity, astronomical waste, marginal utility, nanotechnology, whole brain emulation, and population ethics
* Complexity, universality, and the application of algorithmic information theory to the factorization of probability distributions concerning intelligent agents
* Confidence, confusion, wisdom, rationality, error, and ideology
* Enthusiasm, desire, willpower, curiosity, passion, motivation, exhilaration, and contentment
* Evidence, reason, intuition, and authority
* Personal identity, life, death, risk, and anthropics
* Social identity, belonging, kindness, charity, and secular ecclesiology
* Species identity, evolution, modernity, mental augmentation, galactic colonization, eudaimonia, and hooglamp
* Intelligence, creativity, search, insight, design, and engineering
* Language, meaning, communication, interpretation, conceptualization, and ontology
* Math, physics, philosophy, and cognitive science
* Mechanisms, incentives, laws, norms, judicial precedent, casuistry, contracts, bargaining, negotiation, coalition formation, markets, auctions, and civilization
* Romance, sexuality, affection, commitment, love, and mind-lust
* Truth, possibility, existence, modality, time, space, simulation, recurrence, and thermodynamics
* Value, preference, pleasure, pain, utility, sacredness, wonder, beauty, aesthetics, personality, virtue, neuro-economics, and volition
* Credit, blame, status, respect, reputation, ego, conscience, moral luck, necessity, supererogation, sacrifice, grace, rights, obligations, moral truth, moral inference, and moral progress
* Meta-mathematics, paradox, consistency, reflection, self-reference, and self-modification
So maybe I should focus on nouns like those and forget the magmatic bees. What I'd really like to make is a corpus of Xenant aphorism concerning those topics, with some strange constraints for the sake of alien poetics, like that an aphorism is better if no word appears in it twice. On Earth, brevity is valued in aphorisms, but not so for Xenants, and I take that to be an approachable design challenge. What semantic features are central to aphorisms once you get rid of brevity? I'm working it out, but in the meantime, here are some short aphorisms, almost all from acquaintances on twitter, that suggest the tone I'd like to capture, if not the length or clause structure, in Xenant aphorisms:
A game is a fantasy world simulation with access to human oracles.
A mob is what happens when people of questionable loyalty have to collectively decide what it means to be loyal.
A neutral posture is one from which many actions can be initiated.
A norm is a societal tool for defeating the individual will.
A prophet is a person who writes a small to medium number of sentences featuring suspicious capitalization.
An opinion is a belief backed with a weak weapon. A decree is a belief backed with a strong weapon.
Domains policed primarily through shame will eventually be dominated by the shameless.
Extremist fundamentalism is a fixed-point of the parody operator.
Humility is the ability to trust others to solve problems for you. Patience is humility with respect to time.
Intimate proficiency grows from using stable tools.
Mathematics is hard to generate and easy to verify. Philosophy is easy to generate and hard to verify.
Norms against pettiness exist to keep defections from spreading between games.
Not every error should be called folly.
One has only so much time and energy to decide which battles to fight.
Patience is a virtue, but time is a resource.
The best way to criticize a world is to build a better world.
The longer the delay between action and feedback, the harder the skill is to learn.
The set of fantasy world simulations is the same as the set of computer programs, but the probability measure is different.
The trick to arguing well in person is to have already argued well on paper.
The true name of any god is the source code that it would take to summon it.
Tradeoffs too dreadful to imagine will be decided without awareness.
Your afterlife is the long-term effects of your actions as they pertain to your goals.
Pretty alien, right? Way better than human proverbs like "A fool and his money are soon parted.". A lot of them are almost definitions, and yet, unlike the famous definitional aphorisms in Ambrose Bierce's "The Devil's Dictionary", they're not merely snide insults. It's wonderful. We just need to make them long and disordered and rambling like Cicero, with occasional entomological, crystallographic, and metallurgical references, and we'll have something that we can translate into a new language and then pretend it came from that language originally for some reason.
Truly, I'm excited. For the future of this project. Goodnight.
-
Day four. Ooh, ooh, ooh! You don't need to number the nouns in that scheme where I put numbers on all the nouns and verbs! The nouns can be implicitly numbered by the order that they're mentioned in the sentence. And numbers could even be referenced in verbs before the corresponding nouns show up, for that dangling suspense and resolution pattern, like "I knew that you killed someone, but I didn't know that that someone ... was our son."
Here's the dog sentence again: "The dog who chases the chicken likes the cat who hates any dog who chases a chicken".
We could now write the dog sentence with an ordering like "1,4-hates the-cat 5,2-chases 4,3-chases the-chicken 5,1-likes a-chicken any-dog the-dog", and the inferences in the Xenant's mind as it reads that sentence would be like, "Hate is the cat's preoccupation, as something goes out and chases. It is a chaser of chickens the cat hates, and now something chases after this one chicken, some unknown thing with affections for the cat. Any dog the cat hates, for any dog will chase a chicken, and this dog with its affections, chasing this very chicken, is no exception." The Xenants get all of that suspense and resolution, rapidly and compactly. Good shit. That's the beauty of a very free word order.
The numbers on the verbs are still a little ugly, but I would find them less ugly if I just hid them in Xenant sounds, like "ki'tupi,zi'tiku-chiti'chu". I'll keep trying to think of a better way to enable a free word order, but for now, I can work with this. The scheme effectively provides an infinite class of relative pronouns that are incorporated into verbs for very precise poly-personal agreement. Sounds conlang-y to me.
I think that if nouns and verbs can be arranged freely, then I'm okay with other parts of speech being less free in where they occur relative to those. But I'm not sure yet how I want to do that in order to keep sentence meanings unambiguous. I have some reading to do. Like, I don't know how to deal with non-finite verbs or matrix predicates or adpositional phrases as adverbial adjuncts or most dependent clauses.
Let's take it slow. I don't think I want to put numbers on adjectives to mark what nouns they modify. In English, adjectives post-positional are funny to me for some reason, so maybe adjectives will follow nouns in the ant language, if I use adjectives. If I think up a way that adjectives could go on either side of a noun unambiguously, that would be fine too.
How do I deal with prepositional/adpositional phrases in a language with a free word order? Consider "He hid the note (in-my-pocket) (at-the-restaurant)." The word "put" is a trivalent verb (and perhaps a ditransitive verb, depending on whether you consider an adpositional phrase to be an object) that has an object corresponding to a noun phrase with a thematic role ("the note") and a prepositional phrase with a locative role ("in my pocket"). That's a problem because I've been numbering noun phrases implicitly, not prepositional phrases. I didn't even realize I was numbering noun phrases, I thought I was simply numbering nouns. Oops. And then the last prepositional phrase (at-the-restaurant) isn't an argument of the predicate "put" at all, and is instead called an adjunct, and here it has an adverbial function. How are numbers for marking verbal agreement supposed to be implicitly assigned here? I have to put them on bracketed groups of words. And the brackets could contain their own numbered noun phrases. How awful.
Here's a similar sentence with adpositional phrases and with a definitely ditransitive verb: "He gave me the-note (with-the-information) (in-the-restaurant)". The verb's second object corresponds to the noun phrase "(the note with the information)". But if you don't assign numbers well to groups of words, then when you reorder the sentence, even keeping the bracketed prepositional phrases intact, then the restaurant might seem to have the information rather than the note. This is pretty sad to me. I will write about something more cheerful right now and come back to the issue of free word order another day.
Let's talk about verbs! I would like verbs to be a closed class in the Xenant lang. I've given some thought in the past to making a conlang with just "exist" as a verb. That feels like too much burden to put on this language when we've already got a very free word order (hopefully), but a small set of verbs is still highly attractive to me. I think "be", "do", and "have" makes for a nice set of core verbs, along with a causative operator for upgrading the predicate valence ("cause-Y-to-be-Z", "cause-Y-to-do-Z", "cause-Y-to-have-Z"), and perhaps we could add "can" and "may" as modal auxiliary verbs to reduce how often we have to explain counterfactuals on DAGs and speech acts determining the contours of social decorum (if the ants even have such a thing as speech acts and decorum). Mmm, nah, I don't want auxiliary verbs. I'll inflect verbs for some combination of tense, aspect, mood, and evidentiality, and the mood inflection will cover for the absence of "can" and "may".
I'm very open to the idea that there could be second-order verbs defined in terms of the core verbs. Maybe that doesn't meet your conception of verbs being a closed class, but I'm more interested in closed semantics than closed lexicons. I also like the idea that second-order verbs should come in antonymic pairs by virtue of having each definition include a polar choice somewhere like,(more | less) or (newly | no longer) or (in to | out of) or (toward | away from). It seems like most verbs in English already kind of have antonyms, but there's some wiggle room in which word someone will consider to be the antonym, because it's not clear where in the unwritten definition you're making a polar choice. But if you just write the definitions, and write them two at a time, then you can have opposite verbs, and thereby organize and compress your lexicon. And if don't believe that your human concepts are amenable to clear rigid definitions that way, well, that's okay, because this language is suited for analytical aliens with large working memories, fully capable of posing definitions and then sticking to them.
Okay, today was a little bit of a step back with the prepositional phrases as verb arguments, but that's just correcting for an earlier misstep. Onward we shall continue, on surer footing. Goodnight.
-
Day five. Yesterday, I was wondering about how to number prepositional phrases and noun phrases as arguments of verbs. Maybe I could just not use verbs that take prepositional phrases as arguments though. That sounds fine. And then, instead of numbering noun phrases, I could either avoid ever making noun phrases by using polysynthetic inflection to make big weird nouns that incorporate all the modifiers I want, or I could stick to a more strict word order within noun phrases (but still move noun phrases around in the sentence as a whole willy-nilly for the sake of poetry). And if I keep noun phrases, then when a verb references e.g. the 7th noun, it'll be understood that it's referencing the noun phrase that has the 7th noun as its head.
Okay, both of those options sound kind of bad to me. I want a little bit of inflection and a little bit of a rigid word order within noun phrases, and the phrases will move as units.
I like agglutination more than fusion as a scheme for inflection, but the words can get long with agglutination, and that feels kind of self-indulgent in a conlang. However, if antlang syllables have the same length as a single consonant, perhaps I could do an agglutinating language that still had short words?
Let's find out. First, let's talk about nominal morphology. Xenant nouns don't inflect for gender, case, or animacy. Xenant nouns do inflect for a kind of countability/number, and the options include mass and twin. And maybe they also include nought, singular, and plural, I'm not sure. But it would be cooler to just use mass and twin. The unmarked form of a noun indicates the mass number, which means that the speaker is not currently conceptualizing the thing as having boundaries that would enable counting. The twin number is only used for small numbers of intimately related objects, like two crystals that have grown as twins from shared lattice points, which happens to be how twin siblings grow for the Xenants. Xenants take antonyms very seriously in modeling the world, and they use the twin grammatical number for binary categories, like when referring to "the-truth-values", "the infinite time spans" (past and future), "the lateral directions", or "the size differentials" (larger and smaller). If I find that I can communicate in this language well enough without having dedicated singular and plural grammatical numbers, then I'll end up calling the mass number "non-twin". Xenants also optionally inflect their nouns for definiteness. I don't know what affixes to use for any of these inflections. Probably I'll use those syllables I mentioned before with mixed lateral height on a single articulant (like Oo and xX) that won't appear in the semantic roots of the lexicon.
That's inflectional nominal morphology. There's also derivational nominal morphology. I've got loads of derivational strategies collated last year and just waiting to be associated with morphemes, although I'd like to identify and use a small and parsimonious set if possible.
Just below the level of morphology, at the phonology-semantics interface, I think I'll also have the nominal roots organized with phonesthemes to hint at ontological features like "abstract/concrete" and "endurant/perdurant" and "animate/inanimate" and "sortal/non-sortal".
How about verbal morphology?
For inflectional morphology of verbs, we already have that number-based polypersonal agreement thing, which is basically pronoun incorporation. I don't think we need to inflect verbs to agree with any little feature of the nouns when we've got that kind of ridiculous direct precise reference. I do want to mark something of tense, aspect, mood, and evidentiality on the verbs. Last year I came up with a cute system of (5 * 2) evidential varieties and I still like it. I might give it to my xenants who shall also like it. Or I'll rework it a little bit to appeal to their love of antonyms.
Tomorrow. It's late. Goodnight.
-
Day six. I'd like the part-of-speech of any root word in the in the vocabulary to be distinguishable by its phonotactics without knowing its meaning. Here's a scheme: Verbs in the root vocabulary have only half-syllables (i.e. the syllables are silent in one mouth while the other mouth makes a consonant). Root verbs tend to be three to four syllables in length, but that's not required for distinguishability. Root nouns start and end with full-syllables (consonants in both mouths at once). Suffixes have length two and they consist of a full-syllable followed by a half-syllable. Prefixes also have length two and they consist of a half-syllable followed by a full-syllable.
So if you see hfhhhfh, you know that you've got a verb with a prefix and a suffix (hf-hhh-fh). If you see hffhffhfh, you've got a noun with one prefix and two suffixes (hf-fhf-fh-fh).
Before, I was thinking I'd put the argument-pronoun-numbers before verbs as prefixes, but there are a limited set of prefixes in two-syllables and there isn't a limit on integers, so I'm going to do something else for that. Still deciding.
Adjectives and adverbs aren't in the root vocabulary, but they can be made from other words by morphological derivation, i.e. (root_noun - derivative_suffixes) could be an adjective. Probably. They might get roots if this turns out to be too hard for me to use for composition.
I think this system for part-of-speech phonotactics is the best thing I've done so far for this language since deciding it would have consonants from two mouths at once. This feels so right. It gets me like half of the way to generating words, which had felt like a very distant accomplishment. Now I just need to work on the ontological phonesthemes and root word generation will be about as constrained as I want it to be.
But before I do phonesthemes, I want to be more sure of my consonant inventory and phonotactics. Those two stop consonants that I added as an after thought need some more attention.
Tonight I want to look at what other parts of speech the ants might have, and then do some more work on their syntactical grammar.
What determiners exist in the antlang? There are no articles, because definiteness is directly marked on all nouns.
There might be interrogative determiners in the antlang, but Xenants don't use questions. Instead of asking "When will it happen?", they say something more like "I don't know the time it will happen", or "It would please me to know the time it will happen", or "Your life depends on telling me the time that it will happen".
I don't have any cultural or grammatical reason to ignore other kinds of determiners. Demonstrative and possessive determiners are fine. Can I treat those as adjectives? That would be nice. Likewise with quantifiers. And then distributive determiners will be pronouns? Which won't be a separate class from nouns? I'm not sure that will work, but that's my first thought of what I'd like to try.
What about interjections? Expletives are fun, and that's a problem. Xenants don't have fun. Their emotional range is very limited compared to humans, and they don't have any concept of sacredness or profanity or taboo. I think we'll probably just skip other exclamations on similar grounds. I'm not sure about greetings or response particles or volitives. For the moment, let's say no greetings, some response particles and volitives, but not human ones.
How about adpositions? This is an important choice, and I'm not going to rush into it. I need to do more reading about languages without adposition. Adpositional case marking on nouns is an option, but I don't want case markers. Still reading.
However I represent adposition, I'm pretty sure that once I have it, I can get by without having conjunctions as a separate word class. Conjunctions are great for making long complicated sentences (which is what Xenants do best) and they expose logical and causal and discursive connections (which is also ant-like) but I think I can do both of those things with just adverbs (to describe relationships between clauses) and adpositions where conjunctions would join nouns together. For example "I want milk and coffee" could be "I want milk beside coffee" or "I want milk, I want-ALSOly coffee"). The case of "I want milk xor coffee" is harder. Perhaps that could be "I want coffee not with milk, I want-BUTly milk not with coffee"? Kind of dumb, but it technically works.
"Why not just add another word class for conjunctions, James? If you're worried about the distinguishability of parts of speech, that's not a real problem. You haven't done anything with 1-syllable words, and if you're using hf- and -fh for affixes, then you still haven't done anything with shared-articulant mixed-lateral-height syllables, and you wanted those to have a grammatical function anyway."
True. Very true. I do still have the option of using shared-articulant mixed-lateral-height syllables. And I'd been avoiding 1-syllable words in my mind, because when I only had 3 articulants (6 consonants) and silence, the space of 1-syllable words felt very small (27 syllables, of which 3 were shared-mixed). It felt so small, that it seemed too important to use. But if I add in those two stop consonants, giving 5 articulants (10 consonants) and silence, I've got 65 1-syllable words, of which only 5 have mixed lateral height on a shared articulant. That's a big set. I can fit adpositions, conjunctions, and determiners in that set, and still do something else with words that begin with shared-mixed syllables.
But that's too English, isn't it? It is. So not all of those are going to be word-classes in the antlang. Not all.
And it's late now. Hopefully I get more done tomorrow. Goodnight.
-
Day seven. I want a small and well-structured verb inventory for the antlang. I mentioned structuring with verbal antonyms before in this post, and I did some significant work months ago trying to put verbs into antonymic pairs (specifically, the verbs that label classes the VerbNet). That was good work, and I will revisit it eventually. Before revisiting it, I want to try making a small verb inventory another way. I want to start with a small set of verbs and see how many common English verbs I can represent in terms of one of those.
For example, lots of verbs can be represented as a kind of giving. Instead of equipping someone, you can give them equipment. Instead of considering something, you can give consideration to it. Often you just nominalize the verb and give that corresponding noun, which is fine by me. Here are 80-ish verbs that can be represented as a kind of giving, simply through nominalization (feel free to skip):
. absolve (give absolution to)
. accuse (give an accusation against )
. add (give an addition of _ to)
. admit ('give an admission that)
. allege (allegation that)
. allow (allowance for)
. analyze (analysis of)
. announce (announcement of)
. apply (application to)
. appoint (appointment to)
. approve (approval of)
. argue (argument that)
. assign (assignment to)
. assist (assistance to)
. assure (assurance to)
. cite (citation of)
. claim (claim that)
. classify (classification of)
. compare (comparison of _ against)
. confirm (confirmation that)
. consider (consideration to)
. credit (credit to)
. declare (declaration that)
. defend (defense to)
. define (definition to)
. deny (denial that)
. describe (description of)
. designate (designation to)
. devote (a devotion to)
. diagnose (diagnosis to)
. display (display to)
. dispute (disputation of)
. educate (education to)
. employ (employment to)
. equip (equipment to)
. exhibit (exhibition to)
. explain (explanation of)
. express (expression of)
. feed (food to)
. forgive (forgiveness to)
. govern (governance to)
. guide (guidance to)
. help (help to)
. impose (imposition to)
. indicate (indication that)
. inform (information to)
. inject (injection to)
. injure (injury to)
. inquire (inquiry about)
. interfere (interference to)
. interpret (interpretation of)
. introduce (introduction to)
. invest (investment to)
. invite (invitation to)
. judge (judgement to)
. justify (justification of)
. maintain (maintenance to)
. offer (offer of)
. oppose (opposition to)
. organize (organization to)
. pardon (pardon to)
. permit (permission to)
. praise (praise for)
. predict (prediction of)
. present (presentation of)
. promise (promise to _ that)
. propose (proposal to _ that)
. protect (protection to)
. protest (protestation to _ of)
. punish (punishment to)
. recommend (recommendation of)
. reinforce (reinforcement to)
. remind (reminder to)
. report (a report to)
. resist (resistance against)
. satisfy (satisfaction to)
. secure (security to)
. suggest (a suggestion that)
. translate (give a translation of)
. treat (give a treat to)
. utter (give an utterance of)
Easy! There are some other verbs for which GIVE can do most of the semantic work but they don't have such obvious nominalizations. I think "gift" arose as a Germanic nominalization of "give", and similarly "tale" archaically nominalized "tell". So "telling someone that X" could be written as "give someone a tale that X". It sounds a little weird, but it works. And if "give" can do the work of "utter" (as mentioned above) and 'tell", then I think "whisper" should be in the same class. The sentence "give me a whisper of his name" sounds weird but understandable, and I'm specifically here to make some weird understandable nonsense. To compete with someone could be to give them a competitor, and to fight with them could be to give them a fight, I guess. To compose something might be to give it composition. And to create something is another way to give it existence, so that's definitely a kind of giving, but "give it creation" is clearly wrong, whereas "give it composition" is just a little iffy. Oh well. It's still a give-verb, even its nominalization doesn't play nicely like the others do.
I think "yield" is a give-verb, but it feels out of place with the others. A corn plant can give a yield of corn cobs, so that works. But what about when you yield control? Are you still giving a yield? I've said yield too much in my head and have developed semantic satiation.
I don't like using deverbal nouns in the -ing form as nominalizations, for some reason. But without them, I'm stuck saying that when you teach someone, you give them a "lesson" rather than a "teaching". And likewise, when you warn someone, you give them "notice", I guess, instead of a warning. When you show something, you give it... I don't know. A showing.
When you release something, do you give it release? Kind of. You give it freedom. "Release" is too often used to mean "emotional relief", and mental states are not part of the verb's semantics.
When you open something, do you give it an opening? I guess so, but that's gross. Let's try to put "open" in a different verb class.
To build or make something could be to give it existence and form? When you sort something, you give it order. There are lots of other give-verb that if I were to present them here, I'd want to argue what's being given, but I've mostly satisfied myself that by having the verb GIVE, I can get by with a language that doesn't have a few hundred specific other verbs.
The verb GIVE means something like "cause X to have Y", and I'm probably equivocating lots of kinds of "having" or possession by lumping so many verbs together. I might fix that later. I also might not. It's fine.
"Do you have any other root verbs besides GIVE?" Of course! Here's a small one: I found that GET could cover for a few verbs (acquire, catch, gain, inherit, profit) that are more about physical possession and also GET works for verbs of passive perception (detect, hear, learn, realize, see, sense, remember), in which you're getting sense data. There's also TAKE, which can be used for verbs like kill (take life) and invade (take space) and capture (take autonomy) and destroy, raid, subdue, steal, and cheat. You can take control of things besides people, and then there are less-necessarily-evil TAKE verbs like "use" and "manipulate", although maybe "to use a thing" is to GIVE it a use, like applying it is to give it an application. Maybe it depends on the use (or the language or the species). TAKE also works for active sensory verbs like (check, study, inspect, scan, measure, record, count), from which we derive verb phrases like "take a look" and "take the measure of it", and maybe even "take to heart" for active remembering (whereas passive remembering was a GET verb). I think that TAKE also works for choosing among options (choose, pick, opt, decide, select, elect). And it covers for verbs of making assumptions and determining designations ("take it to be true", "take him as our King"), like suppose, deem, regard, and pretend-that. It works for some economic transactions, like buy and borrow (but not lend or loan). There are some verbs like "strip" and "divide" that are easy enough to analyze as being TAKE-verbs, but on my first pass I put them in a different class for change of state verbs, and I'll probably stick to that. Rid and remove are clear enough. If you escape, are you taking your freedom? Perhaps all eating is taking. Unless you're giving food to your body. Or to eat is to HAVE, but digestion is to TAKE, since your body takes nutrients. If you win or achieve something, is that taking? We say things like "this round is yours". If you sacrifice something, are you taking it from yourself? If you ignore or dismiss something, are you taking your attention away from it? On my first pass at categorizing common verbs, I thought all of these could be TAKE verbs: (find, discover, calculate, compute, derive, determine, estimate, infer, solve, verify), but now I feel less certain. In all of those, you're actively getting an answer for yourself. What more is there to taking than actively getting? I think among those, "estimate" is fine as a TAKE verb. You can take an estimate. And you can take a derivative, but that's just a pun. "Derivative" and the rest should go somewhere else.
I've got more classes and more analysis. I did a lot of conceptual work to make a small but adequate verb inventory. I'm bored of writing about it though. And it's not quite finished. And I need to integrate it with the antonym thing. So maybe I'll just finish it on some other day and then say "the xenant language has only these 8 verbs", and I just won't go through all the details of the inventory's semantic adequacy. No one cares.
I care. I'm no linguist, but I care about concepts and thoughts, and I want to package up some of my insights about them in a nice little passion project, including insights I haven't had yet. So I will. I'll keep at it. Maybe finishing this language is the way to find other people who care about the things I'm putting into it. That's a happy thought.
But I am bored of verbs right now, so I'm going to do a little more phonotactic work.
Once again, our articulants are ⟨ʘ⟩ (bilabial click), ⟨ǀ⟩ (dental alveolar click), ⟨t⟩, ⟨tʃ⟩, and ⟨k⟩. The clicks are not voiced (no "mwha" with the bilabial click), and they're as dry as a bone. There is no spittiness or sucking. Xenant consonants distinguish those articulants into high and low consonant forms, which might be palatized and rounded respectively. Consonants with the same articulant should not appear in one mouth on two successive syllables of a Xenant word. I think that's going to be a little tricky if I want a rich affix morphology, especially where the prefixes and suffixes meet at the nouns' boundaries. Like if I have a nominal prefix that ends in (t,t) with some heights, then this phonotactic rule means I can't have any nouns that begin with a ⟨t⟩ in either mouth. That's a little bit of an inconvenience. I think I can work around it by having different articulants in each mouth for syllables at the start and end of nouns, and at the end of prefixes and the start of affixes. That way even if a prefix ends with (T, K) and a noun stats with (T, K), I can just put the first T in one mouth and the second T in the other.
Let's add in some more rules too! For the challenge! And I'm romanizing the articulants as O, I, T, X, K. I can alternate O with I really fast. I can alternate O against T or X or K well enough, not as quickly, and it tires me. I think I hold my breath when I'm doing clicks, so I get ... uh oh. What's the word? Winded, oxygen-parched, hypoxic. Isn't there another word for that? Out of breath. Pantish. Dyspneic. Maybe I need to get rid of the click consonants before I faint. No, I just need to practice breathing through them.
I can't alternate T and I. It's like a tongue twister. I get dis-coordinated. So Xenant words won't alternate those in one mouth. Maybe I should just stick to one or the other. But they both sound insectoid and I can hear the difference as clear as day. So I'll keep both. I can alternate I with ... uh oh. Oxygen parched again. Dizzy! That's the word. Yeah, this isn't good. I should get rid of the clicks.
The voiceless bilabial click is becoming a voiceless bilabial plosive, P. And I'm getting rid of the voiceless dental-alveolar click. I need another sound though. I don't want F, S, or Th. Sh is just okay. Maybe a voiceless alveolar affricate, Ts? That's not so far from the voiceless dental-alveolar click. And it's very percussive; ts is a commonly used sound when people emulate high-hat cymbals. And I haven't mentioned it yet, but I would like it if I could render audio for very fast Xenant sentences with a drum sequencer. The Xenants are philosophical elites with sick beats, crystalline treats, and six-footed seats. Or something.
Even so, I'm not ready to make this a beatbox language. I'll sleep on it. We'll see tomorrow.
Goodnight.
--
Day eight. Okay! No P in the language, I'm sticking with the bilabial click. And I'm not getting rid of the dental-alveolar click, but I am adding the voiceless alveolar affricate, ⟨t͡s⟩. Ts will be written Z and z in the compact ASCII notation. And I used zi and zu for the dental click before in the expanded notation, but those should be used for Z and z, so the dental click is now transcribed-long as ci and cu, I guess.
I was thinking about adding a voiceless alveolar trill ⟨r̥⟩ to mark numerals. A high trill precedes numerals and is written # in the short ASCII encoding. A low trill marks a decimal point, I guess, and is written ".". I'd previously used a period to mark (half-)silence within syllables, so let's make silence be "_" now in ASCII symbols. The trills are produced in both mouths at once, so we'll just write # instead of ## for the high consonant, and not double up the decimal point either.
The trills are supposed to sound a little bit like crickets. I think if I add aspiration, ⟨r̥ʰ⟩, that sounds even closer. So let's add it. And maybe just a tap would be better than a trill, if the Xenants speak quickly ⟨ɾ̥ʰ⟩.
I had thought about giving the Xenants a bijective numeration system, particularly when I was thinking about Turing machines in natural grammar, or a signed digit system would be cool, but for now, it'll just be a standard positional system.
Hmm, now I need to figure out words for their numbers.
Ooh! Wait! How about a base 6 system, and the numeral consonants all have shared-articulants and mixed-height. Zero through five are written Oo, Ii, Tt, Zz, Xx, and Kk. And then # can terminate a numeral when incorporated into a verb! That way a verb like "3,34-weld" can be written Zz#KkXx#weld".
That feels a little wasteful. If the Xenants are good at making two sounds at once, why are the numerals so redundant? Well, I wasn't sure how to order the language's full inventory of syllables in a natural way to correspond to numbers if I excluded silence (_, _) and the shared-mixed syllables. I can put them in *an* order, but the gaps feel weird. Also, I like the idea that the Xenants conceptualize numerals as nouns (or pronouns) by default, rather than as adjectives, and since the nouns are supposed to start and end with full-sound syllables (not half-silent syllables), that makes for even more spots in the syllable inventory that aren't being used for numbers, which makes the sort of alphabet-order to numeral-order correspondence that I was going after feel unnatural. Also, a good thing here is that by using different sounds for numerals, I allow myself to have fairly short sounds for the small integers and very short sounds for the language's nominal semantic roots. I don't want a situation where OOtt# means some integer, while OOtt with no octothorp means "preponderance" or something. That would just be fuel for superstition, kabbalah, and schizophrenia. Also, this numeral system is not *that* wasteful. I can write 0 through 35 in one or two syllables. That's fine for the use of numerals as pronouns: you're not likely to get more than 35 nouns within a single sentence, even if you're alien Cicero. And if you do, just use a three-syllable pronoun.
Alternative numbering scheme: If I put a high alveolar tap in one mouth, and a regular consonant in the other, I could do base 12 easily enough (0 through 11 being O#, I#, T#, Z#, X#, K#, o#, i#, t#, z#, x#, k#). And then a low tap (written ".") would be part of digits after the decimal point, like (3 * 12^0 + 11 * 12^-1) could be "z#K.", you see. Then in two syllables, I could do 0 through 143 (and also about 2.5 digits of precision for decimals smaller than 1). You could still incorporate pronouns into verbs separated by two high taps (#, #), or a high and a low, or high and silence. Whatever. Oh, I can't have taps in one mouth on successive syllables. Okay, numerals are separated by a (high tap, silence).
Yeah, this scheme in base 12 is better. I can't do a natural ordering over the syllables, but I can do a natural ordering over the consonants. Base 12 it is. And I like that there are now 1-syllable words for simple fractions like 3/4 = 9/12 = "z.". Really good.
In the long-transcription (romanization?) of Xenant words, # and . will be written ri and ru.
While the taps won't show up in the regular nominal and verbal roots, they might appear in derived words. I'm not limiting the taps sounds to one word-class. Deriving adjectives like "alone" from numeral nouns like "one" seems straightforward. And maybe to self-isolate could be "to make oneself be alone". And if you personally can make a verb out of "7", you go right ahead. More power to you. I don't think the taps will show up in affixes though. They'll just be in words for numbers and words derived from numbers.
It would be funny if a Xenant put the name of someone they didn't like at the start of a sentence, and then used a numeral pronoun with a value like 1/12, to mean "that tiny fraction of a man", but Xenants aren't funny like that. Also they don't have men and women. Their reproduction is sometimes like parthenogenesis and sometimes pseudomorphosis. Maybe. I don't know the details of their biochemistry. But I don't think they have biological sex as we do. "Genderlessness and click consonants, James? Are you playing bad conlang bingo?"
Oh hey! The numbers are words. I've made my first words. Congratulations to me.
I think I want to work on verbal affixes next. I like the idea of not having tenses, but I don't think I have the chops to make that language yet. Maybe my next language will be tenseless.
I want a cleaner separation of TAME in the antlang than what you find in most natural languages.
* The language's Tenses are (distant past, past, present, future, distant future), with present unmarked.
* The available grammatical Aspects are the perfective (with the whole event inside the topic scope), the continuous (with the topic scope in the middle of an ongoing event), the inceptive (topic scope is on the start of the event), and the terminative (...). The continuous aspect is unmarked.
* Moods in the antlang are "realis" and "counterfactual", and Xenants make conditional statements in both moods. Taking an example from Judea Pearl, "If Oswald didn't shoot Kennedy, then someone else did." is a realis conditional. A counterfactual conditional is like "If Oswald hadn't shot Kennedy, then someone else probably would have." A mutilated conditional is an inference made in a mutilated causal model featuring a stated assumption as a fixed fact. In the antlang, the realis mood is unmarked.
* I still need to think about how much of that old evidentiality system I want to import.
* Xenants also mark verbs for situational relevance, to indicate whether the speaker is making a generic statement of fact that applies generally or one that applies only in the current circumstance (generic versus episodic). In the antlang, episodic relevance is unmarked. This distinction is sometimes called a grammatical mood, but it's not mutually exclusive with the realis/counterfactual distinction, so I'm not lumping them together. Xenants compose definitional aphorisms the way we compose poetry, which aphorisms exclusively use the generic mood.
With grammatical relevance, the acronym of morphologically-marked inflectional features is now "TAMER", which reminds me of lion tamers. I wonder if the Xenants have their own version of antlions (i.e. ant-eating lacewing larvae).
I also want a causative operator that can upgrade the valence (the number of arguments) of verbs. I think that's inflectional, so its affix will be positioned as a postfix. I never really know what's inflectional versus derivational, unless the affix changes the word's part-of-speech. From "to $verb", the causative operator gives us "cause something to $verb", like how "die" and " kill" are related, or how "have" and "give" are related, or how "see" and "show" are related. This is a really powerful operator that will just about cut my core verb inventory by half. And in principle you can apply the operator more than once, giving verbs like "cause something to kill". The causative operator is applied before the TAMER suffixes. I'm not sure if it should be a prefix or an affix. Prefix.
I also want a negation operator that doesn't change the number of arguments for the verb. It's also a prefix, and semantically it's also applied to the verb before the TAMER suffixes. I think it will mostly be placed closer to the verb than the causative operator in practice, but in principle, you could apply them in the other order for a different meaning, or use them more than once for a more complex meaning.
I think the causative operator and the negation operator are so fundamental that they won't even be 2-syllable affixes surrounding the verbal root with the -fh or -hf pattern. Like maybe I can fit all the "primitive" verbal roots into one syllable (-h-), and then operators will give new verbal roots of greater length (-hh-). For example, if "have" is a primitive root of the language, then from it we could get "not-have" (lack), "cause-to-have" (give), "cause-to-not-have" (take), and perhaps "not-cause-to-have" or "not-cause-to-not-have" if we're getting self-indulgent.
Before deciding on doing that, I had kind of wanted a benefactive operator that could also upgrade the valence of verbs. From "to $verb", this operator gives us "to $verb for something's benefit". You'll have a hard time finding many English verbs that are related in this way, but I thought that my new language should have it, just because I like benefactive and ditransitive verbs. And it feels weird to have the causative operator alone when there could be other operators doing similar work. If you have a verb "to modify", then the benefactive operator gives you "to modify something for X's benefit". So the verb "secure" could become "secure-for-X" and then you can "secure-for the-baby the-cabinet" meaning "baby-proof the cabinet". Or instead of upgrading a transitive verb benefactively, you could upgrade an intransitive verb. For example, "to die" would become "to die-for", which could be realized in a verb phrase as "to die-for the queen". And if you combine it with the causative operator, you might have "(kill him) for the queen' or "cause him to (die for the queen)", depending on the order of application. That's kind of interesting, right?
It is interesting, but pro-social consequentialist intention is *a lot* less ontologically basic than causation or logical negation. If I'm adding "for" as a verbal operator, there are probably a bunch of other things of similar complexity that I should add as operators.
"For" is a preposition, so I looked at the huge list of other English prepositions to see if I wanted to make valence-increasing verbal affixes or operators out of them also. For some reason, I don't care about augmenting the meaning of verbs with prepositions of location or directed motion. Of the remaining prepositions, I did notice an antonym to the benefactive operator: you can do things "in spite of" a person's interest. That's cute, I suppose. It could be a verbal operator. Similar to "for", you can do things "so that X" or "in order to X", and those have a near-antonym in "lest X". I think the complements of all of those prepositions have different syntactic types, so I should rephrase them, but there's definitely some commonality of meaning. You can also do things "with X" or "without X", which could mean "in the company of X" or "by use of X". You can do things in the manner of X (or "as an X") or in a manner unlike X. Two more antonymic prepositions: You can do things "because of" the influence of X or "despite" the influence of X. I feel like I should have some justification for using or not using these antonymic pairs of prepositions as valence-augmenting verbal affixes or operators. But I don't have any justification. So I'll sleep on it. Still remaining on the list of prepositions are some relative temporal reference: before, after, when, while, until, since. I think all of the prepositions referenced in this paragraph are conjunctions, if not coordinating conjunctions. That feels weird to me. I clearly don't understand the relationship between the two categories. Guess I've got some reading to do. And I've got some design choices to make.
There are three classes of adverbial prefixes that can prefix antlang verbs. They are
* Frequency: never, once, again, rarely, often/repeatedly, sometimes/occasionally, always/constantly, initially, finally.
* Quality of Completion: unsuccessfully, almost, barely, partially, totally, exceedingly.
* Telicity: intentionally, accidentally, compulsively, thoughtlessly, under duress, under illness, under chemical intoxication. (Chemical intoxication matters a lot more to ants with pheromones than it does to humans, which is already a fair amount of mattering.)
None of those prefix classes have to be realized before a verb, and none of their values are assumed when not realized. Or maybe "once", "totally", and "intentionally" are assumed. That seems normal enough. But if a leaf falls, it's not doing it intentionally. Maybe "thoughtlessly" should be the default? Most of reality doesn't have thoughts. Or maybe I could add a telicity like ballistically, automatically.
The prefixes aren't applied in a set order. To "almost(intentionally(see)) means to seek, whereas to intentionally(almost(see)) is... I don't know what that would be - entering a place just after someone leaves so that you don't see them, maybe. To intentionally(barely(see)) refers to situations like looking at just a part of something or looking at something through fogged glass. Oh wait, "seek" and "find" feel like more of a semantic pair than "seek" and "see", even if "seek" is an archaic frequentative of "see". What's the relationship between see and find? Seek and find don't imply vision in modern English: you can seek and find blindly. If you learn the location of something from a book, have you found it, or does finding suggest more direct perception of a thing? Hm. Work to do. Almost(intentionally(X) is a powerful sequence of affixes, whether or not "seek" is represented by putting "see" or "sense" in for X. The sequence functions like "try-to": try-to-have = collect, try-to-kill = hunt, try-to-learn = study, try-to-make = develop, try-to-beat = compete-against.
I've heard that "seek" and "find" are paired verbs with imperfective and perfective aspect, respectively. I don't think that's right, but I also don't know what the semantic relation is instead, so maybe. I'll think about it more.
I should also talk about class-changing derivational affixes. For making verbs, there's not much to it. You just take every verb in your verb inventory and make an affix for it. If you have HAVE as a root verb, then make an affix that turns a $noun into a verb meaning "to HAVE the $noun". If you have a causative operator for upgrading verb valence, then you can add that and now your verb is "cause X to HAVE the $noun. If there's a basic verb BE, then you can make an affix that produces verbs of the form "to BE $noun. Or you can use a few BE affixes specialized for different kinds of nouns if you want to distinguish "being in a state" from "being in a role" from being anything else. Basically I need to settle on a verb inventory and then I'll have an inventory of class-altering derivational affixes for turning nouns into verbs.
But we were talking about affixes that go on verbs. So what about turning verbs into nouns?
Cleaning up the derivational affixes in that old post from this blog, I think this is a pretty good set of glosses for the noun-deriving derivational verbal affixes I want to include in the language:
* (v, n): the process of $verb-ing.
* (v, n): the event/state of having been $verb-ed.
* (v, n): the event/state of undergoing $verb-ing.
* (v, n): that which $verb-s.
* (v, n): that which is $verb-ed.
* (v, n): that which is used to $verb.
* (v, n): the place of $verb.
* (v, n): the time of $verb.
* (v, n): the product/result/consequence of $verb-ing.
* (v, n): the cause of $verb-ing.
* (v, n): the art of $verb.
It could be a little tidier, or more symmetrical, or more ontologically rigorous. Like "having been $verb-ed" includes tense and aspect information, which should be in the verb already. Likewise with "undergoing $verb-ing". What about "tending to"? Maybe that affix's gloss is incorporating information already contained in the "frequency" category of adverbial prefixes, particularly the "occasional" value for that dimension. If I want to say that a volcano is in a state of tending to erupt, I can instead say that it's in a state of $verb-ing, where the verb is (frequency=occasionally)-erupt. That's good. I have stricken the entry.
So, it's a pretty tidy, set, but I admit, "art of $verb" feels out of place. I don't know what similar things it should live among. Maybe "practice", "discipline", "philosophy", "religion", "genre"... Those don't feel ontologically basic. Some of them are in the category of "descriptions", perhaps. I'll think about it more. I can clean this up.
I think that's enough conlanging for tonight. Goodnight.
--
Day 9. I was thinking about what the ants would use for personal names, if anything. They could have names derived from ordinal numbers like how the Romans had Primus and Octavia and so on. "Have you seen The Eighth-One? I think it left its phone in here." I think that having ordinal names fits very well with a culture that uses numbers as incorporated pronouns in synthetic verbs. And they could have other names too. "Mica Slave-maker". "Agile Aphid-milker". "Alpha-radiation Child-of-projectile-weapon".
Today was a bad day for me. That's all I can do. I hope your day was better. Goodnight.
--
Day 10.
I was thinking about stacking conjunctions on verbs to keep the order of clauses free. Consider the sentence, "If you miss my train, then you will still know that I am gone, for you can hear the whistle blow a hundred miles away if you listen." We could prefix the verbs like this: "if-miss then-will-know then-that-am-gone for-if-listen for-then-can-hear". A re-ordering of that like "if-miss for-then-can-hear for-if-listen then-that-am-gone then-will-know" can be decoded readily, even if it's not as quick to read. Nice. Although my verbs are already getting kind of long. Maybe I could just precede verbs with compound conjunctions as a separate word. Or something. But I like the idea of stacking up conjunctions somewhere.
I'm not sure how I want to implement the function of adjectives if they're not in the core vocabulary of verbal and nominal roots. I don't want adjectives to be verb-like. That doesn't work with my verbs being a tiny closed class. One option is that I could make a small number of affixes that turn nouns into adjectives. For example, to express "red cup", I could write "tomato-colored cup". Another option is that I could have general nouns like "red-thing" allowing for appositional expressions like "red-thing cup".
I love apposition; it makes for nice wandering sentences like those of Cicero. But I care more about maintaining a free word order for this language than about reproducing Cicero. So I think I'm going to connect an adjective-thing to "cup" by a copular construction like "the cup is a red-thing", since I'm already using pronominal incorporation in verbs to free up the primary word order. That could look like "1,2-is the-cup a-red-thing". You can still put "red-thing" next to "cup" if you like apposition, but the sequential placement isn't determining the compound semantics.
Hm! Or what about "the-poodle a-teacup share a-size."? "Size" is more general than "small", like "color" is more general than "red", so if I use this construction, I can have a few nouns for dimensions instead of having lots of nouns for values of dimensions. Yes, good!
"He wants the brown poodle" could be rendered as, "He 1,4-want some-ochre a-color the-poodle 2,4,3-share", with ochre being a particular brown thing. It's not pretty. But it's mostly not pretty because you're saying numbers, and those feel like nonsensical insertions rather than ordered nominal references. "Some ochre, a color, the poodle, they share" sounds perfectly poetic when you let "they" stand in ambiguously for the noun positions. I don't know if there's a way around this ugliness. What if I write ordinals instead of numerals? "He 1st,4th-want some-ochre a-color the-poodle 2nd,4th,3rd-share." That sounds way better to me. Maybe I'll occasionally write it that way, even though the Xenants aren't using a separate ordinal form of numbers to mark pronouns than they use for naked numerals.
The Xenant version of the puppy sentence is also longer than the English version, and that's ugly. There's nothing wrong with long sentences, but you want your sentences to be long because they're full of content, not because your language is bad at expressing content compactly. If I use "is" instead of "shares", the sentence is just "He 1,3-want a-brown-thing the-poodle 2,3-is". That's much nicer. And I could still have "brown-thing" or "red-thing" derive from a less adjective-like noun. Like instead of having the adjective "tomato-colored", I could have the noun "tomato-(colored-thing)". Good.
Okay, to derive nouns for dimensional values, like red-thing, I need affixes for dimensions, like "-(colored-thing)", where color is a (semi-categorical) dimension. I don't want dimensions to be a closed class though, because you can always derive new scalar dimensions from scalar base units. So maybe I should just have one affix with two slots: one for the dimension and one for the reference entity from which the dimension's value is measured. Let's order the parts as "$reference-$dimension-thing", such as "teacup-size-thing" or "diamond-hard-thing" or "dictator-personality-thing".
I think I should add a rule for the phonotactics of nouns now. Before I said that root nouns all start and end with full-syllables (consonants in both mouths at once). If I additionally require that nouns of three syllables or more don't have two full-syllables in sequence, then the constituent parts of these new adjective-like compound-nouns can be parsed unambiguously. For example, hffffhhfhffhfh could only be "hf-ff-fhhfhf-fh-fh", i.e. a prefix, a small root noun, a ridiculously large root noun, and two suffixes.
That's okay. Not great, but okay. I'll sleep on it. Maybe I'll come up with a better system. Or maybe I'll decide that it's not such a bad thing to have a dedicated word for red-thing that isn't based on another more concrete red noun. Or maybe I'll decide that there are only a few dimensions that xenants would want to use to make adjectives out of, and instead of specifying the dimension in the compound, I'll just have a different suffix for each dimension. Here's a shoddy list of dimensions I started putting together before remembering about derived scalar units, which you should feel free to skip:
* origin, location, material, phase,
* age, duration, frequency, distance, volume, length, breadth, width, angle
* mass, density, hardness, strength, elasticity
* intensity, energy, speed, flow rate, power, temperature, pressure, stress, force, charge
* behavior, function, manner, style
* novelty, rarity, purity, value
* crystal point group, cleavage, fracture, habit, twinning
It's not a good set by any metric. It's too small to be useful for a race that has physicists and engineers. And I know I said that the Xenants have larger working memories than humans, but this list still feels too big to have an affix for each one. It just wouldn't be good design. Good designs are based on opposites and symmetries. Any Xenant and their twin will tell you so.
Goodnight, honey.
-
Day 11. The adjectival-noun compounding scheme combined with the copular construction from yesterday can be used to make some limited locative constructions, like "Heaven is a-bed-located-thing". What if I want to be more precise about locations? "The jar is under the bed" could be rendered as "the-jar is a-(bottom-of-the-bed)-located thing", but I don't have a way to say bottom-of-the-bed yet, and also I'd like the nominal arguments of the adjectival compounds to be exclusively nominal roots (and not genitive derivations like (bottom-of-the-bed), for example).
Wait, I have a verb "HAVE". Can I use that to express the genitive above? "The jar is bottom-located-thing and the bed has a bottom" doesn't work. "The bed has a bottom and that bottom shares a location with the jar" works but I don't like it.
There's another way to be precise about locations that I was already contemplating for other reasons. I think I want to have verb-framing in my language, i.e. the direction or location of states/events/processes will be specified on verbs, but manners of motion or existence will be specified elsewhere (if anywhere). Basically, Xenants have verbs "go-away" but not "run", "skip", "crawl".
Let us introduce the verb "exist". The locative/directive frames won't specify contact, so verbs can be "over" an argument but not "on" it. Contact will be specified some other way when we address the expression of manners of motion and existence. The Xenants will have locative verbal frames for over/under, left-of/right-of, ahead-of/behind, inside/out-of, and near-to/far-from. The near-to frame is also used for "beside". It seems to me that adding a locative frame increases the valence of the verb, so the spatial frames must be operators, like the causative and negative operators already introduced. Now we can say "The-jar exists-underly-in-relation-to the bed". In short, the locative/directive frames serve the function of relative spatial adpositions, but they do it as valence-increasing adverbial operators.
All the same locative operators for existence can be used as directive operators for motion. But what about circling around something? I should probably have circular/spherical frames for those. I could just use yaw/pitch/roll, but I think Xenants would have yaw/anti-yaw, roll/anti-roll, pitch/anti-pitch. And maybe those can be used for locatives too, like "the rope exists-yaw-to the branch". Mmm, no. The Xenants will have the option of doing that, but they mostly won't. The six spherical operators will just be used as directive frames.
What if you're in a gravitational field, but the relative object doesn't have a front/back/left/right? Then you still have yaw and anti-yaw directions, but roll, anti-roll, pitch, and anti-pitch aren't distinguishable. Tricky. I'm going to ... not think about that for a while.
Spatial frame operators let us express some specialized forms of the verb "put". The sentence "I put some-nopales in my-burrito" becomes "I cause(in(exist)) some-nopales my-burrito".
Is it redundant to have spatial verb-framing in my language and to also have those locative constructions from adjectival compound nouns? Maybe a little, but I mostly do different things with them (adjectival versus adverbial, "red-colored-thing" versus "cause-to-exist-over").
Now that I have mostly taken care of the spatial adpositions, I'll feel better lumping together most of the remaining adpositions, as subordinate conjunctions, with the coordinating conjunctions. Nice.
I haven't done any work on the language's script in a while, and there have been some changes to the phonology since then. Let's review.
The consonants in the root vocabulary of nouns and verbs are:
O o (pi pu): the bilabial clicks ⟨ʘ⟩.
I i (ci cu): the dental clicks ⟨ǀ⟩.
T t (ti tu): the voiceless alveolar plosives ⟨t⟩.
Z z (zi zu): the voiceless alveolar affricates ⟨t͡s⟩.
X x (xi xu): the voiceless postalveolar affricates ⟨tʃ⟩.
K k (ki ku): the voiceless velar plosives ⟨k⟩.
There's also a sound for making numerals:
# . (ri ru): the aspirated voiceless alveolar tap ⟨ɾ̥ʰ⟩.
All of those are distinguished into high and low forms, which humans might be able to emulate by palatalization and rounding, respectively.
The Xenants might also use these sounds non-linguistically.
Sh : the voiceless palato-alveolar fricative ⟨ʃ⟩.
S : the voiceless alveolar sibilant ⟨s⟩.
I want a script that combines two consonants into a syllable, and which also relates the high and low forms of each articulant. I'm not sure how the lateral placement of consonants should alter the glyphs. It should be a small change, if any, since it doesn't alter the word's semantics, but it should also be represented somehow since it's a major phonotactic constraint used to speed up pronunciation. Finally, the script should be something that you could carve in stone. Lots of carved scripts have arcs and circles though, and I have no idea how they do that, so maybe "carved" is not going to be a very productive design constraint for me given my ignorance of stone carving.
...
It's coming along really really well! I'll show you when it's done.
The problem with doing my adjectives as "$reference-$dimension-thing" is that the speaker and the listener have to both know of a thing with the right property. To say "lost" you have to have a reference object that's known to everyone to be lost. To say that you're "sorry", you need to reference a thing that's known to everyone to be sorrowful. That's not great. That could use some work.
I realized that if I can incorporate indexical pronouns into verbs, I can also incorporate indexical proverbs into conjunctions. For example, if Sharon says "I love you, but I'm lost", that could be written by Xenants as "1,2-but 1,2-love 1,3-am I you lost-thing", where "1,2-but" is counting verbs as they appear and "1,2-love" is counting nouns as they appear. This is monstrous, is it not? I've created a monster. But it feels like a logical extension from what I've already made.
That doesn't give me conjunction stacking, though, which I still think I want. I'd have to let the indexical proforms for conjunctions range over verbs and conjunctions as they're mentioned in the sentence. *deep inhalation* This is awful.
Is the whole language awful? Yes...
Goodnight.
-
Day 12. This morning I realized I could have indexical proforms range over (phrases headed by) every word in the sentence, without restrictions like "verbs have noun phrases as arguments" or "conjunctions have verb phrases as arguments". That's very regular syntactically, isn't it? At the cost of abandoning semantics. I hate it. I hate it so much that yesterday's scheme of incorporating indexical pro-verbs into conjunctions doesn't feel ugly in comparison.
I think I want conjunctions to operate exclusively on clauses and not on noun phrases. There will be primitive conjunctions that are one syllable each, and maybe the syllables can be stacked together into multi-syllabic compound-conjunction words. I think that by including conjunction stacking, I can have the indexical pro-forms for conjunctions just range over verbs, rather than verbs and other conjunctions.
What will the conjunction look like phonotactically? Because conjunctions will be preceded by numerals (functioning as indexical pro-verbs), they won't be confused with nouns or affixes, which don't get numeral prefixes. So I just need to design the part-of-speech phonotactics for conjunctions to distinguish them from verbs. Verbs are made of half-syllables, and their prefixes also start with half-syllables. So as long as I use full syllables for each conjunction syllable (which can be composed in arbitrary sequence), then I'm not limited in my choices of articulants or phonological heights. That's very freeing.
What conjunctions are in the language? For logical and causal conjunctions, Xenants have OR and XOR, IF and THEN, and AND, THUS, and BUT. I thought about having BECAUSE coordinate with SO across two clauses, the way that IF and THEN coordinate. And instead of having BUT, I could have DESPITE coordinate with YET. Or "Because/So" could be written as "AS/THUS". And I could have EITHER coordinate with XOR. The coordination is very attractive as an illustration of Xenant dualism. It also feels redundant. I could just have IF and not mark the coordinated THEN clause. I'll think about it some more.
For relative temporal reference, Xenants have conjunctions that reflect their tense and aspect systems. The tensive conjunctions are LONG_BEFORE, BEFORE, DURING, AFTER, and LONG_AFTER. The aspectual conjunctions are SINCE (inceptive), UNTIL (terminative), WHILE (continuous), and WHEN (perfective).
I think that's it for Xenant conjunctions. Maybe I'll add some thematic ones later. But this feels really good for now. Hey, whoa, "when" is an interrogative determiner for time. Maybe I should have conjunctions that are interrogative determiners for other things? Like "where" could be a conjunction, couldn't it? But maybe not the other ones? That's a puzzle for another day.
Closely related to conjunctions are relativizers, like "that" in "He knows that (you love him)".
Interrogative determiners can also be relativizers, but I don't want to think about those today.
To make a sentence into a relative clause, I think the Xenants will put a verb-to-noun-deriving affix on the sentence's main verb (and a different affix depending on whether the nominalization of the verb produces an event, a state, a process, or maybe something else). That's all it takes. If the verb has no arguments, then it becomes a gerund, and if it does, then we get a relative clause. The sentence "He understands that the dog is brown." will look more like "He understands the-(the dog is a brown thing)-state".
I don't know if the Xenants ever nominalize a phrase where the main verb has some but not all of its arguments specified. Like if you wanted "give me the duckling" to be a relative clause (not an imperative sentence, just an incomplete relative clause), then maybe the indexical pronominal numerals before "give" would have a zero, to show non-specification of the subject, and also two positive integers indexing "me" and "ducking". I don't really like it, but it does sound useful. And I'm not sure about the gerunds either. But making a clause into a relative sentential argument is solved.
I think I need to do a little more work on making adpositional meanings expressible. I never did anything with that benefactive operator, and I didn't do a good systematic job of putting it into a family of similar things. I'll do that now. Let's look at grammatical cases to figure out what kind of adpositional meanings Xenants might want to express: https://en.wikipedia.org/wiki/List_of_grammatical_cases.
Oh shoot, I just realized that my phonotactics for root verbs are too restrictive. I can't have verbal operators being one half-syllable each and also have as many as I've already come up with, because there are only 12 half-silent syllables that don't use the aspirated taps. I guess I could use the two-syllable affix phonotactics for the locative/directive frames. I had at least ten of locative frames, depending on inclusion of the spherical directives.
Yeah, that's fine. I thought of the spatial frames as being kind of adverbial anyway. Those will just be placed in another adverbial affix class, alongside Frequency, Quality, and Telicity. Spatial Occupation I shall call it, unless I come up with a better name. Habitat? Tenancy? Domain. Relative domain. Idk. Placement.
Anyway, about those adpositional cases. Let's do them tomorrow. Something to look forward to. I've got a guess that the set of case-based thematic operators I settle on is going to look something like "as", "for", "of", and "with". But maybe not! We shall have to find out. Goodnight.
-
Day 13. If I want to augment verbal meanings with non-locative adpositions, similar to how other languages relate nouns to verbs with thematic cases, I might use thematic affixes corresponding to the English "for" like a benefactive case, "as" like an operative case (i.e. a behavioral similative), and "with" like an instrumental case. I'm not sure if I want a similar thing for "about/concerning", as in "he thought-about X". I know that I don't just want to use a locative affix metaphorically, as English does with "about". I think I'll address this short-coming after I've narrowed down the verb inventory.
It doesn't make sense to augment verbs with "of", since "of" related nouns to nouns, not nouns to verbs. I need to think more about how Xenants represent genitive relations. I don't need to worry about genitive possession, since I already have a HAVE verb. For genitive composition, I think xenants use the copular verb: "the-chair is a-gold-thing". And likewise for genitive origin: "the chair is an-Italian-thing". So I need to have nominal affixes that turn gold the substance into "a gold-composed-thing" and Italy the place into "an-Italy-originated-thing". I probably already came up with those and many more in that old post about derivational affixes. I just need to change their types from (n → adj) to (n → n). I also want to make sure that Xenants can say things like "book of laws". Which ties back into that about/concerning relation from the last paragraph! How interesting.
When I woke I up today, I was thinking of other things though! I was thinking that the verb HAVE could have an unspecified number of arguments. The first argument would be the thing that possesses, and all subsequently specified arguments would be possessed entities. Something like (John, A car, A house, A dog, A daughter)-HAS. And I also thought that the copular verb could have unlimited arguments. Something like (John, A student, a young-thing, a father)-IS. But what if I want compound subjects rather than compound objects, like in "John and Mary have a house."? Maybe indefinite-valence HAVE could also be used to form collectives: "(The couple, John, Mary)-HAS and (The couple, a house)-HAS". You could also form a collective this way for the possessed objects, if you wanted to refer to them collectively more than once.
I think manners of motion should be specified as nouns. Instead of "crawling", Xenants might "do a crawl" or "go with a crawl". And nouns can be used by Xenants where English would use change-of-state verbs too. Water doesn't "freeze", it "goes-into a freeze" or "goes-into the-state-of(being(a solid-thing))". Or better yet, getting rid of that metaphorical "into", we could say "starts-to-EXIST-as a-solid-thing". See that "Exist-as"? That's the operative affix in effect. It's already useful.
Oh whoa, wait a second. If I expand the operative "-as" to be a full similative (similar state as well as similar behavior), then "exist-as" can be used for the copular-be! To say, "he is tall", you say, "he exists-as a-tall-thing". Hot damn.
I don't think I have any way to make comparatives in this language yet, like "X is smaller than Y." or "I love you more than anything." or "I'd rather feed you than house you.". That's something to tackle today.
But first, let's talk about nominal affixes.
I don't think the Xenants will have affixes for (current | future | former) $noun. That meaning is encoded in verbs through tense and aspect. Likewise a (wonderful | typical | terrible) $noun will be encoded using the Quality of Completion adverbial affixes on verbs. I'm not sure about augmentatives and diminutives. Those could be good.
Here are some nominal affixes that I think I want, gathered up from that old affix post, cleaned up a little, and organized semantically:
Resemblance (non-behavioral):
* (n, n): a thing with the (shape | size | smell | sound | taste | texture | weight) of $noun.
Origin:
* (n, n): a thing made of/from $noun.
* (n, n): a thing made in/by $noun. A thing originating from $noun.
* (n, n): a thing transmitted by $noun.
Function and purpose:
* (n, n): a thing used passively with $noun. Especially, a thing made for holding/supporting $noun.
* (n, n): a thing made for use by $noun.
* (n, n): a thing that works on/with $noun.
* (n, n): a thing that makes $noun.
* (n, n): a thing that monitors or regulates $noun.
* (n, n): a thing that seeks $noun. A thing that works to acquire $noun. A thing that collects $noun. A fan of $noun.
Genitives (Possession, Pertinence, Parthood, and Participation):
* (n, n): a malady affecting $noun.
* (n, n): a part of $noun.
* (n, n): a member of $noun.
* (n, n): a participant of $noun.
* (n, n): a thing contained in/by $noun. A thing located in/at $noun.
* (n, n): a thing that pertains to $noun.
* (n, n): a thing that has $noun. A thing containing, filled with, or laden with $noun.
The genitive list is still kind of a mess. For one thing, "malady" shouldn't be on the same footing as "part" and "participant". Even if I replace it will with the more general "quality", so that they're on a similar level of abstractness, they don't feel related. I could just say, "a thing affecting $noun". That's... good, actually, maybe.
Also, among the genitive affixes, other than "pertaining", all of their functions can be served well enough by the verb HAVE. The arm has an ache, the tree has a trunk, the group has an accountant, the jam session has a saxophonist, the apple has an apple-parasite, and the worm has a worm-home.
And as for the "about/concerning/pertaining to" concept... that's not very general, is it? It's just symbolic/linguistic/communicative expressions that pertain to things, isn't it? And perceptual experiences. And conferences and fan clubs. Okay, maybe it *is* kind of an important meaning to be able to convey succinctly. I might get rid of the other genitive affixes, and do something else with "pertains to", like make it a valence-increasing thematic adverbial affix. Or something else, depending on the verb that I end up adopting for referential situations "believing that", "saying that", "seeing that", and "feeling that".
When I was organizing the old nominal affixes, I realized that they fell into basically three groups. Some just had an adjective, like "big $noun". Some were indefinite, like "a member of $noun" and "a thing that makes $noun". And some were definite, like "the location of $noun", "the theory of $noun", "the art of $noun". I still need to clean up that third class. I don't know why some of them are definite and some are indefinite. It feels deep. Also there was one uncountable affix: "concern for $noun". Very interesting.
Some of the definite nominal affixes were the (existence | location | time period | state | event | process | quality) of $noun. Those are all nice abstract categories that could go near the root of a Top Ontology, but I'm not sure how productive they would be in actual language use. Like, a sneeze is an event; do we need to say "the event of a sneeze" separately from "a sneeze"? I don't think so. So maybe I should remove state, event, and process. "Role" and "object" and are categories that have a similar abstractness and similar uses as those other abstract things, but I don't think they would make for useful affixes either. Maybe "role" would. I dunno. "Category" and "instance" also fit around here, I'd say. I think "the category of $noun" could be useful. And I'd say that "instance" is what we already have an affix of definiteness for. Nice. Existence, location, time period, role, category, and instance/definiteness. That's pretty good. Time, space, and existence feel like a tidy trio. And roles, categories, and instances are considered a tidy trio in the field of formal ontology. This is good.
There were some other less ontologically basic definite affixes that I sorted through: the (art/practice | behavioral standard of being | purpose | science | study | theory) of $noun. It's kind of a mixed bag, but they're also all... system of thought and conceptualization that people build up around objects of repeated contemplation? "Religion of $noun" and "philosophy of $noun" could probably go in the same bag, but the Xenants wouldn't use those affixes. What about "the style of $noun" or "the system of $noun" or "the expression of $noun"? There are so many words that could fit into that slot. And I still don't know what really distinguished these affixes from the indefinite ones. I kind of want to put "purpose" with "cause of $noun" and "effect of $noun", not that I've put those anywhere either.
Oh! The glosses for the indefinite affixes almost all had the form "a thing that $verbs ($preposition) $noun", like "A thing that is exists in $noun". The only ones that weren't like that were "part of $noun", "member of $noun", and "participant of $noun". That feels important. And parthood, membership, and participation are pretty basic concepts in formal ontology and closely related. People make logics about those things. So maybe that's important?. Hm. This is a tiny bit of conceptual progress. I hope it leads to more. If I tidy up the less ontologically basic definite affixes, I'm things will become clear.
Let's talk about comparatives now!
The first thing to check is whether any of my language features can already express comparatives. I have an idea for this! To say "X is smaller than Y", you could say "if X is small, then Y is not small". It's perhaps more colloquially idiomatic than literal, but I like it a lot. Another pair of coordinating conjunctions can work almost as well: "As X is small, thus Y is not small". I like the IF-THEN version a lot more though. I think the adverbial affixes for Quality of Completion could also be fruitfully applied here. Remember those? Verbs can be done (unsuccessfully, almost, barely, partially, totally, exceedingly). With those we could say, "If X is a small-thing, then Y is-barely a small thing." That's nice because it keeps us from saying "Y is not small" in cases where we think of Y as being small in some absolute sense (perhaps relative to the body proportions of the speaker), if not small relative to X.
I think temporal conjunctions could also be used here. "When X is small, Y is-partially small." But I don't like that. It sounds too much like X and Y and changing size dynamically and inter-dependently.
By using different Quality Of Completion affixes on the copular verbs for X and Y, you could make starker comparisons, like "If Mike is-exceedingly beautiful, then everyone else is-unsuccessfully beautiful." Kind of impolite to say so, but Xenants will be Xenants, and you should see Mike. Mandibles for days. Or maybe they'll just keep the copular verb bare in the superior case and mark it "-barely" in the inferior case. That's fine.
Does the conditional idiom work to express a comparative like "I love you more than anything"? I think a Xenant would say something more like "If I love you, then I love-barely everything else." I think that works. I actually haven't introduced any way to say "everything else" or "everyone else", but we'll get there eventually.
How do Xenants say "We invited more people than wanted to come."? .... Hm. I wonder if the Xenants will have a hard time using the conditional idiom to express arithmetic statements like "5 > 3"...
Okay, first off, I think we need to be able to express a simpler phrase like "the number of people". Then I can say, "the number of people (that we invited) was greater than the number of people (that wanted to come)", and I can put that into the conditional idiom readily.
Nothing comes to mind immediately for how to say "the number of people", but here's how I'd express "Three people exist.": "The-people exist and the people are a-thing-of-size-3". I'm deriving that last noun by an affix of type (numeral → noun) that turns an integer like "2' into a noun like "pair" or "doublet" or "binary-thing".
But what about "the number of people"? Ah! Maybe this is like "book of law". The number is about the people, as the book is about the law. It's a descriptive genitive. My old nemesis, "concerning", has shown its face again. I had been thinking about making "concerning" into a valence-increasing thematic verbal affix. But what verb would it go on in "the number of people" Do I have to use the copular verb, "the-number is-concerning the-people"? That sounds dumb. How about I just have "concern" as a primitive verb? Then I can try to say "the number that concerns the people" by relativizing "concerns" with a verb-to-noun deriving affix. That feels better. And then instead of saying "He believed that X" and "He feared that X" and "He hoped that X", and lots of other verbs that I don't want in my tiny closed verbal inventory, instead I could say "He HAD a (belief | fear | hope) that CONCERNED X. "Concerning" wasn't my nemesis at all. It was my friend, all along, showing up to try to help. What a beautiful moral to this tale. But I might rename it to REFER.
Unless I can do without it. I do want a very very small verbal inventory. And I did also mention wanting a genitive affix that means "a thing that pertains to $noun". I could just say "the number is a-thing-that-pertains-to-the-people" or "the number is a-(the-people)-descriptor" for short. Is that cheating? Am I maintaining a small class of verbs by shoving new verbs into nominal affixes? Possibly. If you look at the glosses for the indefinite nominal affixes, I mention verbs like (affect, collect, contain, monitor, originate, pertain, regulate, seek, transmit, work). Maybe I need to express all of those glosses using just Xenant verbs? Or I need to figure out which ones I can't express using Xenant verbs, and then use that to improve the verb inventory, and just get rid of all the indeterminate affixes.
For right now, I'm okay with having some extra verbs in the glosses for indefinite nominal affixes, since those are also a small closed class. Though honestly, I don't like that the "pertaining" affix would be on its own if I removed the other redundant-seeming genitive affixes. I'd like it if "pertaining" were in a group of other affixes with similar meanings, each of which I had wanted to put in the Xenants' language for their own reasons. But maybe I should just lump "a thing that pertains to $noun" in with the function/purpose affixes. Describing is a function, is it not? Yes, it is. But maybe I should look for some other abstract / informational functions to go with it. These concepts should be presented in their natural contexts, for the sake of regularity and completeness and symmetry and beauty and adequacy and all the rest. And I can always handle the meaning of "pertains" some other way if I want to. This is open to revision.
Temporarily allowing for a "pertaining" affix, I think we have all the pieces we need to represent "We invited more people than wanted to come." It's going to be a little ugly when translated, with relative clauses in relative clauses, but that's a part of speaking complex thoughts. I think I'll translate it tomorrow though. Goodnight.
--
Day 14. I'll get to the translation shortly.
First, I was thinking about determiners. The Xenants won't have possessive pronoun determiners like "my" and "your". More importantly, I think I want to change how I mark nouns for count/mass and definiteness so that determiners are very semantically regular.
If I have a mass grammatical number for substances like "gold", I can have:
* An indefinite and a definite article: Some, The.
* A small and large indefinite quantifier: Little. Much.
* A small and large definite quantifier: None-of-the. All-of-the.
I think that's very nice. What other grammatical numbers can this for using English words? Plural works:
* Articles: Some. The.
* Indefinite quantifiers: Few. Many.
* Definite quantifiers: No. All.
For the singular grammatical number, there are articles:
* Articles: A. The.
But we only have "no" as a small definite quantifier. Perhaps "one" could be used as a large definite quantifier. I'm tempted to use diminutive and augmentative suffixes for small and large indefinite quantifiers, like dogette and dogzilla.
But we're not done. There are distributive determiners, which I want to first analyze as applying to singular things. "Any" dog and "some" dog feel like a pair. And "no dog" and "each/every dog" feels like a pair. And maybe the pair of "no" and "each/every" is a small definite distributive and a large definite distributive determiner. Maybe "any" and "some" are also large and small in some sense, even though they refer to one dog? Like "any" is a large set of dogs that we could draw one dog from, and "some" dog is a set with one member, but we don't know the member.
Here's another conceptual hurdle: I used "some" as the indefinite article for the mass number (some gold, rather than the gold) and for the plural number (some dogs, rather than the dogs). But "some" exists in some kind of semantic pair with "any", and I see now that I could use "any" with the mass and plural numbers also. Is "some" actually indefinite? Is "any" also indefinite? Are there two indefinite articles for plural and mass grammatical numbers? What's going on here?
Finally, what about "neither, either, both"? Those sound like quantifiers for a dual grammatical number. Maybe neither and both are (small/large) definite, while "either" is a lone indefinite determiner. But while English might have dual quantifiers, we don't have a dual number. We're stuck saying "either dog" (singular) and then "both dogs" (plural). It makes perfect sense that there's only one indefinite determiner for the dual number, but that also messes with the nice symmetry of large/small indefinite and large/small definite. And then English doesn't have articles for the dual case at all. But I could just say that the Xenants do have them. Indefinite and definite articles: "A opposites", "the opposites".
I think I need to figure out why English seems not to have a dedicated indefinite article for plural and mass numbers, but it does have a definite article for them, and it can use the distributive indefinite determiners "any" and "some" on them, but it can't use the large definite distributive determiner "each/every". And if there's no good reason for it, then the Xenants might get a more regular system of determiners than English has, and then the Xenants' grammatical numbers will probably be Mass, Singular, Dual, and Plural.
But first, let's translate that sentence from yesterday. The source sentence is "We invited more people than wanted to come.". A more ant-like expansion of that is "The number of people that we invited was greater than the number of people that wanted to come."
Maybe the language can't express that yet. I don't have a way to say, "the people that we invited". The relative clause scheme that I made before just let's me say things like "he saw (the people were invited)-event".
I can say "The number of people was greater than the number of dogs.". That would look like "the-number is a-(people)-descriptor, the-number is a-(dogs)-descriptor, the number if-was a-big-thing, the number then-was-barely a big thing", but with indexical numerals instead of syntax, of course. That's pretty wordy. What if I use the numerals? That would look more like "number people-type 1,2-is number dog-type 3,4-is big-thing 1,5-if-was 3,5-then-was-barely.". No, I didn't use proforms for the if/then conjunction, I just stacked the coordinated conjunction parts on to their respective verbs. It's better that way. Anyway, that's sentence is not too bad. It's not too good either. And it's still not the sentence from yesterday with the invitation.
I spent most of my free time tonight working on selecting vocabulary roots, which doesn't lend itself well to being written up. I guess I should spend more time on making the invitation sentence expressible and the dog sentence short.
If I settle on a system of quantifiers, I could say "We invited some people but few people wanted to come". That's ridiculously simpler. Maybe I should think of shorter ways to express "the number of people" and "the number was greater than the number". That would help with brevity. I still need to be able to say "the people that we invited". I can't say "we invited the people and the number is a people-descriptor", because nothing in the language would link "people" to "people-descriptor", and I'm not ready to have nominal affixes like "-descriptor" attach to indexical numerals. So... if an indexical numeral on a verb refers not to a nouns but to the noun phrase that has that noun at its head, as I've said a few times and keep forgetting, then I just a way to make "the people that we invited" into a noun phrase, and then "the number is a people-descriptor" should be interpreted as "the number is a (people-that-we-invited)-descriptor. So...
Oh, wait, wait, wait. Can nouns participate in more than one noun phrase? If so, then I can't let pronominal numerals refer to noun phrases by indexing the head noun. I mean, it hasn't been a problem yet because I haven't had any noun phrases; determiners attach to nouns, prepositions attach to verbs, and relative clauses have only been sentential verbal arguments so far. But if I'm going to say, "the people that we invited", then I might have to be clever about things. Like, if I want to say "the number of people that were all lonely that we invited", I have to make sure that the bracketing / dependence structure can be reconstructed.
This sounds hard. I'm going to go think about roots for a while. Maybe something brilliant will come to me in the middle of the night. More likely it's going to take some time and some reading about how natural languages with freer word orders usually handle relative clauses and complement clauses. Maybe Nahuatl! As a treat. It's always nice to learn a little more Nahuatl.
I expressed comparatives and superlatives with a conditional. Should I express equatives with a conditional? A while ago I thought about using a verb "share" that could express equative statements, but I think I abandoned that for copular constructions and adjectival nouns. Maybe... maybe I should re-read this whole journal and take stock of what I've done well and what I've done. Tomorrow.
Ooh, I can do equatives using the similative verbal affix! "It is as big as a horse" can be translated as "it is-as a-horse a-big-thing". I like that.
Good work, James. Goodnight, reader.
-
Day 15. Oh, I absolutely did have relative noun phrases already, I just forgot about it. "The dog who chases the chicken likes the cat who hates any dog who chases a chicken". I just need to take that pertinence affix off of "people" so that people can be referred to by multiple verbs. I could put a pertinence affix on "number" instead of people. But I still need a way to link "pertaining-number" to "people".
Using spatial metaphors, I could say "The number has-over the people pertinence". But if I told the Xenants that, they would wonder what it means for a number to have pertinence under the people, and I have no answer for them. Likewise, if "the-number exits-inside the-people", then what does it mean to exist outside of them?
If I use the instrumental affix, I could say "the number exists-by-means-of the people". And the people are kind of used to construct the number in the process of counting them. But let's keep looking. Even better, I could use the operative affix (behavioral similative) to say "the number exists-as the people". That's good because it suggests a dynamical logical relationship: when the people change, so does the pertinent measure. However, if I use that for pertinence, I need a different / more complicated expression for "book of law", because the book doesn't logically change when the law does. I'd have to say, "the book exists-as the author's past conception of the law" or something. But maybe that's good; I need to be able to distinguish between realizations of measures (like the author's current conception of the law or like published population statistics) from unrealized properties (like the state of the law or the number of the people in a population satisfying some predicate). So maybe logical references will "exist-as" their referents and things like beliefs and books will "exist-about" their referents. That's pretty good.
For ontological nominal phonesthemics, I think I'll just have some ontologically distinct top categories as two syllable nouns (ff) and longer nouns will make those more specific by putting syllables inside. Like "decision maker" could be a category that speciates into professions and hobbies and roles in games and things, maybe. "Organism" could be a two syllable noun that speciates into biological taxonomic species. "Artifact" could speciate into hand tools, complex machines, appliances, vehicles, buildings, weapons, clothing, containers, supports, physical documents, furnishings, sensors, transmitters, fasteners, toys... And that's not a very regular list, so I'm probably missing some things. Maybe I need to better organize my concept of what artifacts exist. But I like the idea of general nouns speciating with infixes. It would be really cool if the infixes were meaningful, so that "decision-making" and "organism" could speciate in the same semantic direction when given the same infix. That's probably not something I can do, but I'll think about it.
What does the space of possible infixes look like? What is the set of strings to which we could apply speciating meanings? To keep nominal roots distinguishable when surrounded by affixes (hf-, -fh), I previously required that nouns of length 3 or more wouldn't have repeated full syllables. That means speciated nouns will internally have sequences of half syllables separated by lone full syllables. Like f( hhh f hhh f h)f. It makes a little sense to me to group half-syllable runs onto the full syllable that precedes them, so that speciated nouns will necessarily have a run of half syllables (which you may remember is also what verb roots look like) and then optionally a bunch of fhhh... objects, that I will call... some ducks, because they have a bunch of half-sized things following after a single full-sized thing.
I could actually relax the requirement of not having consecutive full-syllables inside of a noun, because I later further required that the nouns start and end with mixed-articulant full syllables. So you could have two full syllables inside of a noun so long as at least one was a full syllable with a shared articulants in each mouth. But let's keep going on without that phonotactic relaxation for a second.
Now, it would be cool if the speciating infix that looked like a verb root (the train of half-syllables) actually was interpreted as a verb root. Like the verb root for "not exist", when used as a speciating infix, could mean a fictional thing (decision maker, organism, artifact...) and the root for "have" could be a collector of things as a decision maker, or a container as an artifact, or a viviparous organism (a species whose young develop in their bodies), or something. The root for "go" could speciate an organism to mean that it was motile, and it could also speciate an artifact to make it a vehicle (or an object in generalized class that includes vehicles for conveying people and also projectile weapons as vehicles for conveying destructive force). And "exist-concerning" could speciate artifacts into physical documents. Pretty cool, right?
But then, if you're interpreting hhh as a verb, there will be a natural inclination to interpret -fh ducks after the verb root as a verbal affix. So now we can't assign new meanings to half-syllable runs or to -fh ducks, and any meanings that we want specifically for use in nominal speciation will end up relegated to the duck-families of length three and longer like -fhh and -fhhh. That's not ideal.
I don't know that I'm going to stick with speciating infixes being verbal, but I definitely want to try it. Although like, what if you want to refer to an organism that's motile and viviparous? I've heard of such creatures. How would the Xenants do that? And also, it's it kind of awful to be forced to use descriptive phrases referentially, isn't it? Language users should be able to assign short codes for objects of common importance. You should have a short word for "antlion" if antlions eat your people every day, and not just call them "the terrible brown motile things" or whatever. And meaningless names help a lot with making short codes. Is there a way to combine the two?
The space of two full-syllable words is quite large, even if they're both mixed-articular syllables (which I required for the phonotactic-purpose of keeping the ants from speaking the same ariticulant on successive syllables at the point where affixes meet nominal roots). Since the space is large, maybe I could have some ontological nouns that speciate verbally, and some other nouns of daily-importance in 2-syllables whose forms don't indicate their ontological characters. But I don't like that. It would be better to have meaningless infixes than to have meaningless 2-syllable nouns. Or I could just have infixes that aren't interpreted verbally, despite looking like verbs. Or...oh! I'd forgotten how big my consonant inventory has grown. Two full syllables with mixed articulants gives me more than four thousand words. I could do fine-grained ontological types on the first syllable, have the second syllable be arbitrary, and still readily refer to antlions in a way that told you they were organisms. That changes everything. I don't need complex infixes for nouns roots. Suppose I let myself optionally use a 1-syllable (half-syllable) infix, of which there are twelve. What are the 12 regular ways that I'd want to speciate nouns across ontological classes? That's an interesting thing to think about. I'll think about it if those four thousand noun roots ever seem inadequate.
I think human-made lexical ontologies usually don't give enough design consideration to the concept of roles as distinct from sortal categories, and I'd like the Xenant language to do a better job on that front than we do. Maybe there will be a way to make roles out of sortal classes, so that you can see from the word "mate" that it's a role that organisms can generally perform (with each other), while "steed" is a role that's specific to motile organisms, and "lap dog' is a role just for dogs. I'm also a human and do poorly when thinking about roles, but I'll give it the old Harvard try.
I had a germ of an idea that I could use ordinals for comparatives instead of using conditionals. To say "Lions are bigger than sheep" as a Xenant, you'd say something like "lions exist-concerning size first, sheep exist-concerning-size second". Maybe "exist-concerning" is the wrong verb, but there should be some verb that could be used with ordinals to make comparisons out of conjunctions.
A brief bit of word building: I've said that Xenants have a limited emotional inventory. Let's flash that out. I think they don't have separate motivation systems for liking and wanting. They can feel urgency and they can be attentive. They can be surprised, but behaviorally it mostly looks like a pause. They do track reputation in iterated games in a way that allows for them feeling trust and betrayal, but they don't have an emotion of hate for someone who consistently attacks them. They don't have a sense of pathogenic disgust. They don't have fun, humor, profanity, or sacredness. I don't think they feel lust for potential mating partners, but maybe, depending on how I decide that they reproduce. They feel regret if they did something stupid but not sadness over loss. They don't make pained noises when damaged. They don't have ecstasy and particularly they don't turn into drooling puddles when they get something that they really like/want. They do have a sense of curiosity and boredom, but they have very different aesthetics from us of what's interesting and what's boring, partially due to having different sense organs and different rates of mental processing, although your aesthetics would still be different from theirs if you had their organs and their mental speeds. They don't get jittery in excited anticipation. I think they have some kind of fear, but I'm not sure how it manifests. Generally, emotions for them mean a change of mental priorities, and not much of a change of bodily function.
I haven't addressed Xenant personal pronouns yet, as distinct from indexical numerals that function as verb-incorporated pronouns. I'm sure the language doesn't need personal pronouns as a separate part-of-speech; I've heard that Japanese pronouns aren't a closed set, and they inflect like other nouns. But I'm not sure we even need Japanese-like pronouns. Instead of "I", we could say "the speaker". Instead of "you", say "the audience". If you want to talk about a different speaker or audience, then you just specify something about where or when they exist. And if I decide on having proximal / medial / distal demonstratives, then the Xenants could say "this speaker" versus "that speaker". Easy. And I've already talked about making collectives for verbal arguments, so that just about covers plural pronouns.
I ran my aphorism finder on this page. "Perhaps all eating is taking.", "Good designs are based on opposites and symmetries.", "Language users should be able to assign short codes for objects of common importance.", "Most of reality doesn't have thoughts.", "Meaningless names help a lot with making short codes.", "Fish are just water bees." Who needs a 20k word blog post when you've got 6 short sentences as wise as these?
Tomorrow, I want to narrow down the verb inventory more. I'll find some verbs that I don't immediately know how to express as a Xenant, and we'll work through making them expressible in the language. Goodnight.
-
Day 16.
Verbs of joining and linking like (join link attach combine connect fasten fuse merge tie weld) could be expressed over a plural noun or a collective by causing t/them "to exist-as a unified thing", or by "causing them to have some unity". I also think the adjective-like noun "a-unified-thing" should be formed from the noun for "unity". There will be an ontological category for dimensions/qualities/properties like "unity", and some affix to make "unified-thing" and "non-unified-thing".
Some of the verbs I was having a hard time subcategorizing/translating would be much simpler if I had comparative adjectival nouns. Like, "to fortify" is to cause something to be a-more-durable-thing. I could put an augmentative affix on "durable-thing", but that feels more like "a-very-durable-thing" than "a-more-durable-thing". I could put an adverbial affix on "cause-to-be" from the quality-of-completion adverbial class in order to get a meaning like "to-(exceedingly | partially)-cause-to-be a durable thing", or "to-cause-to-(exceedingly | partially)-be a durable thing", but none of those feel quite right either. I could use the dimensional form of "a-durable-thing" and say "cause to have some durability", but that still doesn't clearly suggest an increase in durability over time, I'd say, because you could be taking away from the total durability but leave some.
Similar to using the coordinated conditional conjunctions IF-THEN idiomatically to express comparative declarations about static states, I could use multiple independent clauses (with different tense/aspect) related by a conjunction or conjunctions to express comparatives for dynamic situations, like "It was a-durable-thing and it was-just-now-caused-to-be a-very-durable-thing". That works. It's seems a lot uglier than sticking an affix somewhere that means "newly-more", but it works.
I'd like the Xenants to be good at math. And related to that, I'd like for their language to be able to simply express basic arithmetic, at the very least. And that means that a comparative verb like "multiply" (make larger to a specified degree) shouldn't require two independent clauses and a conjunction. So instead, I'm going to... think about that.
Okay, I'm considering adding a verb "increase" to the verb inventory. What verbs do we already have? I definitely want EXIST and HAVE. I considered a copular-be, but the similative affix "-as" can do its job when appended to EXIST. The verb HAVE could probably be expressed as "exist-with", although that "-with" affix wouldn't be the same as the instrumental "-with" affix that I mentioned earlier. I'm okay with keeping HAVE for now. I've also had in my mind that "GO/MOVE" would be a Xenant verb, but searching through this post, I can't find any time when I justified its inclusion. I mentioned the verb once when talking about speciating nominal infixes, but I didn't, like, show that it could stand in for seventy different English verbs by combing with different nouns and a small set of adverbial affixes. That's weird. But I did talk about directive frames for motion. The verb "move" is the implicit verb in all of that discussion. Should I express "go/move" in terms of something else, like "exist", "increase", or "change"? And if I have "go", "increase", or "change", does that significantly help me to express addition, multiplication, and exponentiation? I hope so.
It also occurs to me that the Xenant numeral system can't express negative numbers yet. That's something to fix very soon.
I didn't get much done today, but it's time to call it quits. Goodnight, reader. Let's do this again tomorrow.
-
Day 17.
I had an idea that would hugely change the morphology. If verbs start and end with half-syllables, and nouns start and end with full-syllables, I could use infixes to express modifications that don't change the word's part of speech, and use circumfixes when I do. It feels like it might make my words much longer, and I don't like overly long words, but I should investigate it. The indexical numerals are prefixes right now, so that would have to change too. They would be infixes. I suppose that would look something like "to-cause-3-to-exist-with-4-into-5". It wouldn't have to be an English-like infix order, but I think it would definitely look very syntactical, which already feels too normal. I wanted an alien language without syntax, not a language that shoves syntax into morphology. I might as well just take a normal human language and then remove the spaces separating words.
Some of the English verbs I was having a hard time expressing as a Xenant have implied-contact. Touch, hit smash, carry - things like that. My first half-thought of how to handle those is to have a noun "contact" and then you can "$verb-with contact". Some verbs like "bring" imply accompaniment but not contact (and also don't imply non-contact of course). And again, my mind leaps to the form "to-$verb-with accompaniment". Some verbs also imply the use of a body part, like "chew", "stomp", "hug", "kiss", and "climb". I think the instrumental affix that I introduced earlier is suitable for that, but again it would be rendered basically as "to-$verb-with the-foot". Are the '-with" affixes in the contact and accompaniment situations also instrumental? If they're not instrumental, are they at least the same kind of -with as each other? If they are the same kind of -with, are there other nouns that verbs can be -with in that way? And is motion one of them? I think "to-exist-with motion" feels a lot like "to-exist-with accompaniment". Are they all just manners? What other manners should I be able to specify? Maybe "damaging force", which distinguishes "hit" from "touch".
What other nouns might go with "contact", "accompaniment", "motion", and "damaging force"? Disorder, control, speed, support, rhythm? Some verbs also specify non-contact, like how "float over" differs from "go over".
The Xenants could say that they cause X and Y to be "contacting-things". Much like comparative adjectival nouns, such as "a-redder-thing", "contacting-things" would be a relational category that doesn't make sense of one object without a relator. I don't like it. I won't do it. The Xenants could instead say that they "cause X and Y to have some contact". Much better. That works for collectives ("X and Y") and plural nouns ("the dogs"). I think it works for singular and mass nouns too, although in those situations you'd usually include an instrumental prefix like "I caused the dog to have-with a-brush some contact". It would be nicer to put the instrumental affix on "cause", like "I caused-with a-brush the dog to have some contact", but "cause" isn't an independent verb for the Xenants, it's a valence increasing affix. Or maybe that doesn't matter? Stick the affixes wherever you want? Maybe.
In English, "I barely touched him" is often contrasted with having "hit him", which makes me think that "hit" for the Xenants could be readily rendered as "exceedingly touch". Or maybe "hit" could be expressed with the instrumental affix like in "the boxer caused (his hand and the opponent's face) to have-(by-means-of) destructive-force some-contact. I don't really like "force" as an instrument, but it's an option. It's certainly not just material objects that can be instruments, because I'd like to be able to say "give-by-means-of clicks a message", but I'm not sure what category includes clicks and hands but excludes destructive-force. Maybe a click is a sound event and force is a property of a process. You can do things by means of objects and events and processes, but not by means of their properties. You can't do something by means of brevity. That might not be it exactly, but I will figure out some semantic restrictions on thematic instruments.
How would the Xenants say "bring"? How about just "he and the package are going-to the party" and we'll leave it for the listener to guess that the two are spatially co-located and maybe in contact. Or "he has the package and he is going-to the party" would be fine. And if you want to say that he's bringing it in his car, you say "he has-inside his-car the-package and (he and his car) are-going-to the-party". Although I don't think nouns are marked for possession, so there would be one more clause about him having the car, but these aren't fully translated sentences. Or you could use the instrumental prefix to say, "The package is going-(by-means-of) him (/and his car) to the party. So many options. We definitely don't need a "bring" verb.
Okay, I still think it's a problem that I can't say "more" in the language without using two clauses with different tense/aspect. Let's add an adverbial class or operator. We've already got (frequency, quality of completion, and telicity), from a long time ago, and the locative frames, and a few thematic affixes like the instrumental and the similative. Also the causative and negative operators are kind of adverbial. How about just adding "increasingly" as an operator, alongside the causative and the negative. We don't need a range of degrees of decrease/increase as we have ranges for the frequency and quality adverbials. We can fruitfully apply the operators in different orders. For example, "to-not(increasingly(have) mass" means to stop growing, while "to-increasingly(not(have) mass" means to shrink. There's also a clear and useful difference in meaning between "increasingly cause X to be Y" and "cause X to increasingly be Y", although I'm not sure I want to express degrees of causation. If you want to, now you can.
And maybe this gives us a means to express numerical addition. Using the instrumental affix, we could say that "to add 3 to 4" is to cause 4 to increasingly-exist by means of 3? No, that's dumb. And "multiply" would be to exist increasingly-increasingly? Nope, still dumb. How about to to add (3 and 4) is to cause (3 and 4) to be a united/summed thing. And then you can have adjectival nouns for "multiplied thing" and "exponentiated thing" and "tetrated thing" and any other hyper-operation you like. That's better.
And for division, you could "cause (3 and 4) to be a fraction". That's good. What about subtraction? "Cause (3 and 4) to be a subtracted thing" doesn't sound right, even though that's symmetrical with the other operators. Cause them to be a difference? Better, but still iffy. We could use a spatial metaphor and say "cause 4 to not be in 3". Ugly. "Cause 3 to not have 4"? No. Cause to exist-using (3 and 4) a-difference? Still kind of weird, but it sounds the best to my ear of the options so far. I'll probably stick with "be a substracted-thing" for the sake of symmetry. Or I could use that last form for the other arithmetic operators too: "cause a product to exist by means of (3 and 4). Or cause 3 to exist-as a product by means of 4. Something like that. I'll decide another day on the exact form, but now the ants can do arithmetic. And instead of having a dedicated sound to represent the negative in -5/12, I could just "make 0 into a difference using 5/12". Good enough.
"Adapt" took me a second to express, but I think that to adapt to X is "to-increasingly-exist-concerning-X-as an-adapted-thing". I would like it better if the adjectival noun wasn't "an-adapted-thing", but I'm still fine with this. Or you could use a pair of clauses, and then "to adapt to the heat" would be "to-increasingly-exist-as a functional-thing despite the-heat being an unhealthy thing". Good.
I previously expressed some sensory verbs with a verb TAKE. Forget that. To taste something is to begin have knowledge of the taste of a thing, to see something is to begin to have knowledge of the appearance of a thing. You don't take the taste of it.
There are some English verbs that involve speech acts, like "christen" and "negotiate". The Xenants don't make as much use of speech acts as humans do. They don't allow things, or forbid them, and they have a limited concept of "accept" and "refuse" that is demonstrated by behavior more than speech. They don't create official names for things in ceremonies. Their version of promising something is to openly set up a situation where they would incur heavy costs if they were to not comply with a plan description, or occasionally to create and share a computer program that will act on their behalf. I'm not sure what contract negotiations before the promises are made would look like for them, but there is conversation preceding bilateral promising. They don't sentence you for your crimes, they just command a party to apply the sentence. They don't declare war, they just start committing acts of war. They don't ask questions or give each other orders, but they make empathic appeals and threats that function similarly. Their ethical system is much closer to "might makes right" than humans are comfortable with, but it's not all impulsive mighty individuals abusing weak individuals: they do have coalitions and bargaining, game theory and reciprocity and reputation tracking, kin selection and functional decision theory, and other stuff that allows for a complex society. But there are definitely also some aspects that humans would call shockingly immoral. That's just how alien species are outside of fiction. I'd like them to have different ideas of ownership than humans, but I'm still working out what. I think probably they don't have a sense of debt, owing, deservedness. Kind of related to speech acts, a Xenants wouldn't tell another what it "should" do; the first one just gives the second a reason to act how the first sees fit.
Okay, back to verbs. To record and to measure aren't expressed with TAKE anymore. You just cause a recording or a measurement to exist-about the subject.
I'm particularly curious about expressing difficult verbs that have senses with three or more arguments. The next verbs I want to tackle are:
1. Those where you can "$verb X as a Y": (treat regard dismiss ignore use manipulate operate handle obey conceiver-of).
2. Those where you can "$verb that (X is the case)": (verify hear read think pretend imagine know assume estimate suppose comprehend infer find discover determine derive calculate compute suspect check prove).
3. Those where you can "$verb to do X": (choose elect opt decide wait pretend try attempt), which interestingly don't include (select pick).
4. Those where you can "$verb with X" in a state of agreement/fit or disagreement/misfitiness: (comply contrast affiliate associate cooperate match differ agree meet conflict).
5 Those where you can "$verb someone to do an $action": (teach instruct coach prepare persuade convince cause force coerce allow forbid prohibit require).
I'll do that tomorrow. I should probably consult Beth Levin's work or FrameNet. There's so much fantastic work on verb semantics there. But I also kind of want this language to be 100% James Gaukroger. We'll see. Goodnight.
-
Day 18.
I don't want move/go as a separate verb. To GO and COME are to EXIST increasingly (near-to | far from) a place. If you want to say that the object is rolling, you can use add a similative (it exists as a wheel / round-thing) alongside the increasing-directive frame of motion, and leave your audience to determine just how similar to a wheel the object has really become. Or better yet, you can introduce manner nouns like "a roll", "a dance", "a crawl", and say that the object moved by means of a crawl.
I think if I make a different "-with" affix in order to express the HAVE verb, I can probably get the language down to having EXIST as its sole verb. That would be so cool. Let's try it. I hereby create a possessive thematic affix.
I'd like the have the thematic affixes arranged in one place. Here they are:
* Benefactive: "for"
* Instrumental: "by means of"
* Possessive: "with"
* Depictive/Respective/Pertinitive: "about / concerning"
* Similative /Copular: "as / like"
I would love some input on what to call the "about / concerning" affix. "Depictive" and "respective" are common words that are already in my browser's spelling dictionary, but I've actually seen a reference to "pertinitive" in the literature and not the other two.
For the locative affixes, (over | under, left-of | right-of, ahead-of | behind , inside | out-of, near-to | far-from), I could probably have just one affix per pair and a negating thing to get the correlative location/direction, but I'd rather have lots of affixes and short words than few affixes and long words. I also don't think I want "left-of/right-of" any more. That's not very useful. Instead let's do "parallel to / perpendicular to". That pair has a little bit of use as a locative frame and a lot of use as a directive frame, like for expressing beside/past/along. Good.
Okay, the 4th class of verbs I put on the to-do list yesterday are easy enough to express. If we had relational adjectival nouns (like a-redder-thing or a-contacting-thing), we could say "to comply with the rule" is "to exist concerning the-rule as a-compliant-thing". Without relational adjectival nouns, we'll say "he caused himself and the rule to increasingly exist with some compliance". And if you want to clarify that he changed his behavior rather than the rule, you could add an instrumental affix "by means of his behavior". I think that same pattern of 1. nominalizing the verb to create a quality/dimension/property (comply -> some compliance), and then 2. saying that the quality is made to exist increasingly in the collective of related things (the rule and the person), that pattern can be used for all the verbs in the 4th verb class mentioned yesterday. To differ is to be a differing thing with respect to X is to cause yourself and X to have some difference. To associate is to be an associated thing with respect to X is to cause yourself and X to have some association. The exact instrumental means might differ for different verbs, but the pattern holds. Good job, James. One down.
Let's tackle class 2 verbs, like "To know that X". I think I've got it! Or rather, it's a heterogenous category, but I've solved the larger portion of it and now I can focus on the smaller portion! Here's the solution for the larger portions. It comes in three steps: First, "to know" is to have knowledge, "to imagine" is to have a mental-image, "to pretend" is to have a pretense, "to assume" is to have an assumption, et cetera. Second, recall how we previously treated relative sentential arguments by lexically deriving nouns representing perdurants (entities like events, states, and processes) from the verbs that head the relative clauses:
"To make a sentence into a relative clause, the Xenants will put a verb-to-noun-deriving affix on the sentence's main verb (and a different affix depending on whether the nominalization of the verb produces an event, a state, a process, or maybe something else) The sentence "He understands that the dog is brown." will look more like "He understands the-(the dog is a brown thing)-state"."
If we just combine steps one and two, we get a floating/unrelated argument. "He has some-understanding the-(the dog is a brown thing)-state". The third step is to introduce the "about" affix! He has pertaining to the-(the dog is a brown thing)-state some-understanding". The only thing this might be missing is that in "he knows that the dog is brown", he doesn't just have knowledge about the situation: he knows that the situation is true. He has positive knowledge, we could say say. I could just use nouns like "positive knowledge", "positive assumption", "positive mental image". Is there another way? Should I be introducing the noun "a-true-thing" somewhere? I'm not sure. Seems like I should. It feels like I've expressed the first half of "He has knowledge about the situation and his knowledge leads him to believe that the situation is true of reality", but now I should focus on unpacking everything after "and".
A few of the class 2 verbs can't be expressed as "possessing a (positive) mental object that pertains to a perdurant". There's no mental object that pairs with: (verify | hear | read | check). And a few other verbs have related nouns, but it feels a little iffy to say that someone possesses them: (discover, determine, find). Does discovering X mean having a positive discovery pertaining to X? Maybe. Sounds a little wrong. Having "a finding that X" feels different from having "a proof that X", doesn't it? I'm not sure how. Let's focus on (verify | hear | read | check) first. We can express these with the copular construction and a sentence-derived perdurant. First the copula: "To verify it" is to make it a verified thing, to hear it is to make it be a heard thing. Those adjectival nouns don't derive from dimensions/properties; I'd call them deverbal rather than denominal. Is that a problem? It might be a problem. Let's stick with them for a second and see how the copula plays with the perdurant. To verify that (the dog is brown) is to... cause the-(the-dog-is-brown)-situation to be a-verified-thing? I think so, yeah. It's no worse than "to cause the situation to be a verified thing". Oh wait, there's a noun for "verify": verification. But it's a mass noun rather than a count noun, right? You don't have "a verification", rather things have verification after being verified. Or, wait, no, *you* have some verification pertaining to X after verifying X. Right? No, that's not right. The thing has verification "in your eyes" or "with respect to you", or something. This might be getting into the semantic territory of the class 1 verbs from yesterday, like "treat X as a Y".
Class 1 Tricky Trivalent Verbs:: I think "to treat something as a Y" is to categorize it as Y for the purpose of an action. And maybe for the Xenants, putting a thing in a category is expressed as causing the category to have a member. "We treated him as a child in our deliberation" is "We caused the boy-category to have him for (the purpose of) our deliberations", using the benefactive affix "for". Or maybe the pertinence and similative affixes can be combined. "He exists concerning our deliberations as a boy". The similative affix is good here because I don't have to say "boy-category". I think I'm using "for" and "concerning" the same way in the previous two sentences, when they should have different uses. I need to look back at how I've used them before: either they have different meanings, and I'm saying two different things above without realizing it, or they have the same meaning and I need to get rid of the benefactive. Oh, maybe there's some ambiguity with the benefactive in causal constructions: for "X to exist for Y" suggests that X has an intention relating to Y, but when You cause X to exist for Y, it's unclear whether "for Y" is your intention or X's. Is there an affix-order effect with "concerning" too?
The treat-as verbs could also be expressed by relating with a temporal conjuncion an "exist-as" clause to a second clause giving the context of consideration. He exists as a boy "when we are deliberating" or "when we consider him" or something. I don't really like it, but maybe it could be combined with something else to make it better.
I can't do any more. Goodnight.
-
Day 19.
This morning I couldn't think of how to express "it existed for three hours". Clearly, first you nominalize "it existed" into a perdurant. Now we can say ""Its existing" had a duration, the duration existed-as a-thing-of-size-3". But we're still missing the unit. Which affix do we use to connect "hours" to a verb? We could stack up the copular/similative: "the duration existed -as some-hours -as a-3-thing". Ugly. Semantically heterogenous. We could say that the measurement is being expressed "by means of the unit hours". That feels right here, but it feels weird when the unit is people, because then you expect the people to have been used as causal agents, rather than as passive count-subjects. Unless you introduce a unit for for counted objects and say, "the people had a count and the count existed by-means-of the-object-unit as a-3-thing". I kind of like that, actually. In Industry, things are sometimes packaged or sold "by the each" when there's no associated unit for a countable thing. This has always sounded dumb to my ear, but the idea of a unit like "each" clearly has its merits, regardless of the poor choice of unit name. Let's look using the pertinitive affix real quick instead of the instrumental: "Its existing" had a duration, the duration exists-about hours -as a-thing-of-size-3". No that's not good. Actually, you could probably stick the nominalized perdurant into the slot for "exists-about" and thereby skip the introductory of "its existing" had a duration". That would look like "A duration existed -pertaining-to "the process existing" -by-means-of some-hours -as a-3-sized-thing". Not terrible. Or maybe we could just say "pertaining to the process" or pertaining to whatever "it" is that has a duration of 3 hours.
Oh wait, this is clashing with a thing I said earlier. I wanted the copular / similative affix to be used for inhering properties ("The number exists as the people") and pertinitive to be used for instantiated measures ("the statistics existed about the people"). Um... that wasn't great though, was it? Maybe I need to separate the copular and the similative into two affixes. The duration exists-as a 3-sized things copulatively, but it exists as the duration similatively? Still not great. Let's keep moving forward with how things were in the last paragraph, and not think about the exists-as and exists-about distinction from a while ago. That can be forgotten.
How long would that actually look in Xenant syllables? "duration" and "it" and "hour" are probably each two syllables, because that's as small as nouns get in the language and they seem basic enough to fit into those 4-thousand-ish roots. They might get affixes for count/number/definiteness. The noun "a-3-sized-thing" is the number 3 (one syllable) with an affix that turns it into a noun, and I'd previously wanted my affixes to be two syllables (-fh or hf-) although I'm starting to rethink that. The main verb for the duration "exist" is just one syllable (h), but then we have to mark it for TAMER and it also gets instrumental affixes, in this case pertinitive, instrumental, and similative (3 * 2 = six more syllables). There's another verb for "it" existing, and that one gets a nominalizing affix. And we still have the indexical numerals. This is getting long.
Oh! I know. The verb affixes will be one syllable, since they're frequently used and often used in sequence. The nouns affixes can stay two syllables long. I don't think they'll get nearly as much use, and certainly they won't form towers of affixes as the verbs already have. Also, all verb affixes will be suffixes now. When you add a suffix, it will modify everything that's come before it. That solves the ambiguity over order of application if we were to put the causative operator on one side of the verb EXIST and the thematic operators on the other. And let's put the indexical numerals after the affixes that introduce them, instead of making them prefixes that are far away from the affixes they're related to.
This is good. "I gave the dog a collar" is now something like "existed the dog, -possessing a collar, -as-caused-by me". That looks long, but in the language it will be more like "h1f2f3 dog collar me". Hm! What about tense, aspect, mood, evidentiality, and relevance? I might have been thinking of the TAMER affixes as applying to the "cause" part of verbs with a causative operator, and not to "exists", maybe? Or maybe not. I hope not. Exist is supposed to be the only verb. I shouldn't be applying TAMER to "exist" and "cause" separately, nor applying it to "cause" and leaving it unspecified on "exist". I think I'm going to require that the TAMER affixes go at the end of a verbal affix chain. I don't want "exist" being future tense and "cause to exist" being present tense, which could be used to mean "cause it to exist in the future". None of that. Temporally distant effects can just be specified in another clause related to the causal clause by a conjunction.
Now that the "cause" affix is on the other side of the verb, it's starting to look a lot like the instrument affix, "by means of". I think it's fine to have both, but I need to be more clear about what each affix does. Let's try a sentence with both, using this new affix ordering. "I painted the house blue with a brush". Maybe we could write that as "existed the-house -as -blue-thing -as-caused-by me -with a-brush"? That sounds just okay. Maybe I need to bite a bullet and say that "cause" is a separate verb, and it causes perdurants to be in effect, and the complex perduants are referenced in the language by nominalizing verbs that head verb phrases. Like "I caused (the house existing as a blue-thing).". That makes me sad, probably because it feels semantically right but it's very contrary to the style I had in mind for the language. And it might mean doubling up on the TAMER affixes.
I never settled on a system of grammatical evidentialities. Let's do that now. *Reviews old system*. Okay, it's actually still completely beautiful to me. If I have a system of marked evidentialities, it's going to look a lot like this. Let's say that the Xenants don't have the "intuitive" grammatical evidentality. They might have intuitions, but they have strong epistemic norms against expressing statements without stronger justification than intuition. And maybe we should combine the deductive and the inductive into one evidentiality, because how often do we reason with one system and not the other? Rarely. What should we call the combination? Cognitive? Constructive? Generative? Ratiocinative? ... Ductive! Perfect. Their full set of grammatical evidentialities is now "perceptive, weak perceptive, ductive, weak ductive, quotative, and weak quotative". The perceptive will be unmarked.
I'm still not sure if there's an alternative option to 1) always using the passive voce in causative existential constructions ("she existed with a locket as caused by me") or 2) making "cause" a separate verb that should probably operate on nominalized perdurants. But let's think more about causal means and causal agents for a second. Lots of things can cause things. AIs, animals, enzymes, tomatoes, tornadoes, cancer, cannons, caustic bases, markets, drugs, pulsars, et cetera. Maybe any of those things can in principle be causal means for another agent. A doctor can use a drug to heal, a god can use a tornado to kill, and an AI can manipulate you to get out of its box. Consequentialist agents can use other causal agents as means. So maybe the specification of instrumental-means in a verb should imply the presence of a consequentialist agent. Non-agents can also be instrumental-means for consequentialist agents. Like we wouldn't say that a brush caused someone's hair to be disentangled, but an agent can still use a brush to disentangle hair. And then it's more like "the process of brushing" directly caused the hair to be disentangled, and the agent caused the process, with the brush as a process participant.
What do verbs that use the operators "not" or "increasingly" look like now that we have this new affix ordering? "The dog did not come to the house" becomes "the dog not(increasingly(existed)) near the house" becomes "((exists the-dog -near the-house) increasingly) not TENSE=past, ASPECT=perfective". That's a minimum of eight syllables. And each syllable has two sounds. So it's basically a word with 16 letters. Oh well. Sometimes agglutination gives you long words. I can deal with that. Anyway, length aside, there doesn't seem to be anything new and unaccounted for semantically when I put the affixes in order like that. That's encouraging.
After thinking about it a little, I'm okay with "cause" being its own verb that takes nominalized perdurants as arguments. It's pretty cool. And if I'm going to have two verbs in my language, it's way better to have "exist" and "cause" than "exist" and "have" or "exist" and "go/move". Causation is very basic in reality, so it can be basic in my language. I'm okay with doing it that way, now lets see if it actually works. How do we say "He put the note in my pocket at the restaurant"? Cause is always going to have two arguments, so I think I'll put the verb in between to separate the indexical numerals. An SVO morphological order is fine. And then I should put the subject of "exists" before that verb too.
"My pocket" is "the-speaker existed -possessing the-pocket". Them "the note went into my pocket" is "the-note began to exist -in the-pocket". To make it a suitable argument for the cause verb, we put an affix on that "existed" to make it a perdurant, like "existed-situation". And finally we can say "he caused existed-situation -in the-restaurant".
We could assign numbers to nouns like:
#1: the-speaker
#2: the-pocket
1-existed, with-2
#3: the-note
#4: (3-began to exist, in-2)!noun
#5: he
#6: the restaurant
5-caused-4, in-6
And then maybe for brevity, just within this paragraph, E is exist, C is cause, Ep is existed in the past, Cp is caused in the past, Epi is existed with inceptive aspect (began to exist). All together now: "He put the note in my pocket at the restaurant" becomes "I had a pocket, the note was in my pocket, he caused it in the restaurant" becomes "the-speaker the-pocket 1-Ep-with-2 the-note (3-Epi-in-2)! he the-restaurant 5-Cp-4-in-6". Honestly, it's not that much longer. The dashes wouldn't be in the ant language and the adpositions wouldn't be written out: those are just here for legibility. The main thing that makes the alien version longer than the English version is replacing "my pocket" with "the speaker existed with the pocket". If I allow for a first person possessive affix and replace the verbal adposition "in" with "I", since it's supposed to be just one Xenant sound/glyph/syllable long anyway, then the sentence becomes "my-pocket the-note 2EpiI1! he the-restaurant 4Cp3I5". That's not bad at all. I'm not going to add pronouns or personal possessive nominal affixes just yet, but they do seem to make things nice and compact.
Holy shit! The derived perdurants could reference themselves! Like "(1-exist)!" at the beginning of a sentence could mean "the situation of existence existing", lol. That's hot garbage. I'm never doing that. And two perdurants won't reference each other either. Except maybe when they're doing mathematics.
Some nouns in the root vocabulary will represent perdurants, so perdurant nouns aren't only derivative. Now that we have CAUSE as a verb, it seems like the ants could say "we caused a-flood" rather than "we caused a flood to exist". Right? Yeah, because you could replace "flood" with its meaning, and the meaning would be a nominalized clause like "some-water existing over everything hereabout" or something like that.
I'm still not super confident in my decision to include the CAUSE verb. It did help us to separately express "exist-in-the-pocket" and "cause-in the restaurant". That's worth something. We can't do without fine-grained adpositional attachment. Although, if I just used a passive voice was a thematic causal affix, we could still have fine-grained adpositional attachment by stacking up locative affixes, now that we have fixed the order of affix application. Hm. Also, I think that the TAMER affixes on CAUSE might be related to the TAMER affixes on the perdurant nouns that they take as arguments. Like ... "I was causing him to sneeze" is more readily interpreted in English as "I was causing him to sneeze repeatedly" than as the literal aspect-mixing situation that's literally written. So maybe some TAMER-TAMER combinations are too weird to be worth expressing, and the restrictions on that will allow me to put CAUSE back as an operator or a thematic affix.
Let me show you some of my glyph ideas. I've never embedded an image in a blog post before...
Success! I want a syllabary for the Xenants. Above you see six syllables, and each one has a consonant coming off of the upper grey line, and a consonant coming off the the lower gray line, and also there's a solid black line that rests above or below. The solid line tells you whether to voice the upper or the lower consonant with your dominant mouth. That way two syllables like (K, T) and (T, K), which are semantically identical in the language, will have the same glyphs except for the line. The high and low forms of consonants are related by a horizontal flip (like a swoop going down to the left or a swoop going down to the right). The top consonant on the left doesn't have a preferred direction, so maybe that represents silence, which doesn't have high and low forms? I was just seeing what looked cool. I designed this horizontally, with an intention to then rotate it to be vertical, but it didn't look as good vertical. So I'm posting it the way that looks good, but I still want the Xenants to write vertically, I think.
Here's another idea for syllabary:
I've got other designs for two scripts that are very ant-like, but I haven't given them a featural syllabic interpretation, so they're just pretty scribbles and dots for now.
Eventually I should decide how the sound are produced in the Xenant mouth. Is there a cup that gets slapped? Is there a ridge that gets clipped? Is there a silicone tongue slapping about? Maybe there are mandibles hitting other organs, and the manner of contact can be fricative or plosive or whatever, and the phonetic names I've been using for the consonants won't have to change much. It seems like the mechanism of speech should influence the writing system, you know?
That's it for tonight. Tomorrow I'll try to finish up analyzing those classes of tricky verbs with three arguments. Maybe having CAUSE as a separate verb will give me some inspiration. Goodnight, pumpkin.
-
Day 20.
I think I'm getting rid of the adverbial affixes for frequency. If you want to say "he went to the store twice", you say "a count exists pertaining to (his-going-to-the-store) as a two-sized thing" or maybe "(his going to the store) was a repeated-thing". If you want to say, "he goes to the store frequently", you say "his going to the store" exists-as a-frequent-thing". And maybe we can use a similar pattern for the Quality of Completion adverbial affixes, using adjectival nouns like "an unsuccessful-thing" "a successful-thing", or, even better, "a-complete-thing" and "an-incomplete-thing". And we can also use adjectives for typicality, like "(quartz being green) is an-unusual-thing". I think if I have that, I don't need the Relevance class from the TAMER affixes (situational relevance versus generic/gnomic relevance).
But let's get back to the adverbial affixes. I think the adverbial affixes of Telicity should be expressed as arguments of the possessive thematic affix. Basically, you distinguish whether the agent causing a situation acted with forethought (of how to achieve the actual end) and whether they acted with choice.
* Forethought (+), Choice (+): Intentionally
* Forethought (+), Choice (-): Under duress
* Forethought (-), Choice (+): Thoughtlessly / Automatically
* Forethought (-), Choice (-): Accidentally
I'm not sure that "accidentally" goes in that last slot, but I don't know what else should, or where else it could go. It might be correct where it is. Anyway, there are no affixes for telicity now.
I've got some ideas for getting rid of the affixes for grammatical evidentiality, but maybe I should slow down on cutting everything out of the language. I was mostly getting rid of affixes because I want the set to be small enough for all the verbal affixes to each be syllable, which is a really dumb way to design a language.
I should work on conjunctions first. I never really decided if they would prefix verbs or stand alone with their own indexical numerals, nor whether they would come in correlative pairs or stand alone in their semantics.
But first those tricky trivalent verbs from three days ago.
Class 5: Those verbs where you can "$verb someone to do an $action": (teach instruct coach prepare persuade convince cause force coerce allow forbid prohibit require). I think this is "cause someone to exist -possessing some-coaching -pertaining-to $action". This pattern of expression has a very similar short-coming as pattern of expression I used for class 2 trivalent verbs (like "know-that X"). The expression "Have a belief pertaining to X" didn't quite specify positive belief before, and now, if you give someone some persuasion pertaining to dancing, that doesn't quite indicate whether you persuaded them to dance or to not dance. It suggests "dance" over "not dance", but I feel like the sentence is missing just a little something. I could just say "positive-instruction", I guess? That's not good, but we're going to move on for now.
Class 3. Those verb where you can "$verb to do X": (choose elect opt decide wait pretend try attempt), which interestingly don't include (select pick). Oh no, same problem. You can have a choice pertaining to X or a decision pertaining to X. I think I need to just stop using the pertinence affix for these. Also some of the verbs in this class don't work even a little bit with the nominalization and pertinence affix, but I'll get to those later.
I kind of don't like the benefactive affix, "for someone's interest", as much as the other thematic affixes and I was thinking about expanding its meaning or getting rid of it. Maybe expanding it will help me to say "decide to X" whereas above I'm saying something more like "decide about X".
For expanding the benefactive affix, I was already thinking about other uses of "for" besides "for someone's interest". We could say "for the purpose of X", but that feel like something the Xenants would say with a multiple clauses and a conjunction. Like instead of "he was waiting for me to pick him up", they would say "he wanted me to pick him up and so he was waiting". Except that doesn't work either, because my first thought of how to express "he wanted me to pick him up" is to use nominalization and the pertinence affix. "he had a want pertaining to being picked up", so maybe the expanded benefactive's meaning can't be achieved with two conjunct clauses that way and it would be good to add it to the language.
Let's try applying it: He had a decision "for the purpose of dancing"? Maybe. I don't think so. I still want to change or expand the meaning of the benefactive, but that particular option for change doesn't help with expressing this one verb phrase. What if we talk about the benefactive interest of things very generally, not just agents? He had a decision "for the interest of dancing"? Or even more simply, a decision "for dancing"? Again, maybe. Again, I think we can do better. Let's go through the other thematic affixes and see if one of them does better than the pertinitive or the benefactive. The instrumental's no good (exist with a decision with/by means of dancing). Nor the possessive (exist with a decision with/possessing dancing). What about the similative/copular? He existed with a decision as/like dancing. Nope. That old "positive decision pertaining to dancing" pattern is starting to look more attractive.
Maybe I'm missing an indicative/quotative affix. He gained some-knowledge indicating-that (the dog was a-brown-thing). He gained a message saying-that (the dog was a-brown-thing). But a message can indicate different things to different people. It's not a fixed fact of a message what the message indicates. X indicates Y to Z. Do I need an affix that introduces two arguments (the interpretation and thhe interpreter)? Or the person who exists-possessing the message could always be the interpreter, implicitly. "He caused her to have a message indicating-(to her) that (she should wear a hat to the party)." Except the xenants don't have the concept "should" like that. Instead it would be "indicating-(to-her) that (he wanted for her to wear a hat to the party). Wait, why is "for" the adposition that attaches to "want" or "have a want" or "exist-with a want"?
The affix "indicating-that" feels a lot less adpositional than all the other thematic affixes, which makes we worry that I'm adding a verb to the language, instead of decomposing verb meaning into (Exist, Not, Cause, TAMER, thematic and locative adpositions, and nouns).
I'll come back to it. Let's work on determiners some more. Nouns will have ontological phonesthemics indicating whether they are countable or massed/uncountable. Counted nouns are considered singular by default. Both massed nouns and counted nouns can take a null quantifier prefix (no-gold. no-dog). There are several more quantifiers that the mass nouns can take, and when they're used on the count-nouns, the count nouns then gain a plural interpretation. These are the universal quantifier (all-gold / all-dogs), the small assertive existential (not-much-gold / few dogs), the large assertive existential (much gold, many dogs), and the indefinite assertive existential (some gold, some dogs).
That's determiners and quantifiers. Do I still need articles? I've been using "a" and "the" all throughout this post. It seems silly to be reluctant to include them. And yet...
I don't want personal possessive pronominal prefixes. But maybe possessive nominal compounds would be alright. "Speaker-hand" is the speakers' hand. "City-hall" is the city's hall. I wonder if... if there's something more powerful I could be doing with nominal compounds though. Maybe somethin where the compound semantic depends on the ontological categories of the component nouns. "Carnival-speaker" would be interpreted according to the even-agent pattern, and "mason-hammer" would be the agent-tool pattern, and "zebra-clam" would be the organism-organism pattern, and "gold-thing-book" would be the material-artifact pattern or something. There would be some rules that would let you know that "zorb-kopf" is a zorb's kopf, and not a kopf made of zorb, or a kopf used on zorbs, or a kopf that works in a zorb, et cetera. Nah, I think it would be better to just keep my nouns uncompounded and represent relations between them with verbal phrases, but it's tempting. Maybe for my next language.
Let's revisit conjunctions! I really really like the idea of stacking conjunctions. Let me give another example:
* (because still) you do not have long,
* (so if) you're going to shoot,
* (because despite-that) you think you have time,
* (so then) do it now,
Present those four in any order, and I can still figure it out. It's beautiful to me. If I were not to prefix verbs with coordinated conjunctions, I would compress them all down to un-coordinated conjunctions, which would stand alone as separate words, with their own prefixed indexical numerals referring to verbs and other uncoordinated conjunctions. Let's see how that would look.
#1: do it now
#2: you're-going-to-shoot
#3: 1-IF-2
#4: you-think-you-have-time
#5: you-do-not-have-long
#6: 4-BUT-5
#7: 3-BECAUSE-6
So much worse to read, right? But it produces short words, and I do like that aesthetically. And it makes sense within the language; why would the Xenants only use indexical numerals for nominal arguments of verbs and not for verbal arguments of conjunctions? If it's worth using once, it's worth using twice. I'm not ready to decide between the two schemes, but I'm glad I've illustrated them now.
I was thinking about what sorts of adjectival nouns the Xenants would have. I think they don't have privative adjectival nouns like "fake, phony, toy, counterfeit, past, imaginary"-thing. Privative when referring to adjectives (rather than adjectival nominals) means "$adjective X" implies that the thing isn't an X. To express "toy bird", they wouldn't say "it exists as a toy-thing and a bird", because toy birds aren't birds. They might instead say "the shape of the toy exists as the shape of the bird" or something.
The Xenants do have manner adjectivals, like "a-skillful-thing", but they only apply them with a pertinitive context, like "he exists-as a skillful-thing with-respect-to surgery". Formally, the English manner adjectives ("skillful") related to the Xenant's manner adjectival nouns ("skillful-thing") are characterized by being subsective but non-intersective, meaning that ($adejctive noun) is a kind of noun, but it's not necessarily an $adjective $other-noun, even if the thing is an $other-noun. Like if you point to a skillful surgeon who is also a cook, he isn't necessarily a skillful cook. That's how Xenant manner adjectivals work. They apply within a context. "Typical-thing" and "bad-thing" are in the same adjectival category as "skillful-thing". They work the same way. "Recent-thing" is also a manner adjectival (non-intersective but subsective), but I'm not sure the Xenants describe nouns with intertemporal relations like that.
The great majority of the Xenant adjectival nouns are intersective. If X is a sick carnivorous thing, then it's also a sick thing and a carnivorous thing. There's no weird interplay. It's so easy.
There also adjectives and adjectivals that are non-privative, non-subsective, and non-intersective. Barbara Partee calls these adjectives "plain non-subsectives", and provides for examples: (potential, alleged, arguable, likely, predicted, putative, questionable, disputed). The noun phrase "disputed bird" doesn't tell you whether it is a bird and also a disputed thing (intersective), whether it is a bird but it's only necessarily a disputed thing with respect to being a bird (subsective), or whether it's not a bird at all (privative). The plain non-subsectives tell you the least of all the classes. The Xenants do not have these as adjectival nouns, and mostly express these meanings with the grammatical evidentiality suffixes on verbs.
I'd love for some part of the Xenant adjectival noun system to be based on the work of the linguist Marcin Morzycki, but I'm just not that good.
Next, I'd like to start deciding on ontological classes for use in the nominal phonesthemics. I need to look up some old work I did on ontology for that. It would be pleasing to me for phonetic reasons if there were just under 30 ontological classes, but I'm going to try to come up with a good set of classes firstly, and only let the phonetic considerations influence the number and representative granularity of classes secondly, if at all.
This is basically going to be me putting all English nouns into whatever groups I think make sense. I'll probably start with perduants. It's going to take a while. Let's call it a night. Goodnight.
Actually, this post is over 30k words. I'm going to start writing in another blog post tomorrow. Conlanging part II.
