Evidentiality in non-humans

A forum for all topics related to constructed languages
Post Reply
User avatar
LinguistCat
sinic
sinic
Posts: 238
Joined: 06 May 2017 07:48

Evidentiality in non-humans

Post by LinguistCat »

So, humans tend to be seen on the whole as a visual and auditory species. Not that we have The Best vision and hearing, but that these are the primary ways we experience the world, when we don't have some form of disability regarding these. Scent is powerfully connected to memory, but we don't have many descriptive terms for scent compared to sight and sound based words. We have a lot about taste but we don't go around tasting everything we run into. Touch (which is actually several senses, but often bundled together) is probably a close tertiary sense, and can also stand in for how we experience emotions, but AFAIK doesn't show up in evidentiality, which is what I'm interested in today.

I think if non-human species had languages that used evidentiality, they'd likely order what kinds of evidentiality were used very differently from us. Most evidentiality in human languages, from my research, seems to go (personal experience/non-personal experience) > (personal/hearsay/other non-personal) > (personal-vision based/other personal/hearsay/other non-personal) > (personal-vision/personal-hearing/other personal/hearsay/other non-personal). And no languages that I've heard talked about(!) separate out personal experience based on other senses than vision and hearing, but I'd be happy to be proven wrong.

But I could imagine, for example, that a lot of youkai in a story world I'm working on would use scent-based evidentiality, and probably even extend that metaphorically as well. A species that evolves under the ice layer of an ice planet or the moon of a gas giant might make a lot more use of some combo of taste and scent and so that might be their main focus in evidentiality, followed by touch. Though I suppose it might be difficult at all to get into the mindset of species that mainly experiences it's world through touch and scent and has never had vision at all (though maybe they echolocate, that could be fun).

Do any of you have evidentiality for non-human languages? What forms are important and which are grouped together as an "other" category. Do they interact with humans and if so has evidentiality changed because human speakers/listeners (or substitute the appropriate senses for the language as well) have difficulty with that form of evidentiality? If so, how has it changed and if not how do the humans work around it?

What kinds of evidentiality systems would you like to make or see made? And any thoughts on what kind of creatures would use it?
Salmoneus
MVP
MVP
Posts: 2232
Joined: 19 Sep 2011 19:37

Re: Evidentiality in non-humans

Post by Salmoneus »

I don't think you've paid sufficient attention to the differences between what we might call recognitive sense and what we might call constructive sense.


Sight is a paradigmatic 'constructive' sense. What I mean by this is that through sight, we receive sense-data that is, in itself, fundamentally meaningless, and from it we (mostly subconsciously) "construct" a "world": we imagine a volume of 'space' and 'time', within which 'events' occur. We then rationalise out 'participants' from that event. So, I look up, and amid a field of blue I see some brown light moving toward some grey light, and then the two colours moving in unison. From this, I construct an event: the capture of a prey bird by a predator, against the blue sky. From one event to another, I can then follow each participant through time and space: the hawk, and the pigeon.

Chemical senses, on the other hand, are recognitive. What I mean by this is that through smell, we receive sense-data that is directly linked to, and directly symbolises for us, the presence of an item. This is why we "don't have many words" for smells - in fact, we have an endless number of words for smells, but they're incommensurable. The smell of an orange is called "an orange smell"; the smell of a dog is called "a dog(-gy/-gish) smell", and so on. Whereas sight gives us a very small number of basic sensation types - red, green, blue, light, dark - from which we can build more complicated building-blocks, from which in turn we can build comparable, commensurable descriptions of things we can see but not recognise, smell instead gives us a vast number of different sensation types, all more or less unalike one another, not composed of one another (while we can draw analogies - X smells sort of like Y - these are famously idiosyncratic, with different observers reporting wildly different analogies, and even the same observer reporting wildly different analogies at different times; famously, for example, the female sense of smell - not only preferences, but analogies and even objective acuity - varies as a result of the menstrual cycle, and through the course of pregnancy). Humans are able to recognise more than a trillion different smells, and other animals can recognise far more than this. Smell directly senses the object.

Smell does not, however, incorporate any inherent concept of space, or even of time: the smell of a thing can remain long after the thing itself has left. [at least, as visual beings would describe it. It's likely that a olfactory species would see the skunk as continuing to be 'present' in some sense so long as the smell of skunk remained - for an olfactory species I suspect objects are less like point-objects moving through space, and more like four-dimensional clouds or fogs, that frequently interpervade one another...] As a result, events - occurances in space and time - are invisible to smell. If a fox kills a hare in front of you, and you have only smell to detect it, then all you actuall sense are three (to simplify!) objects: fox; hare; blood. You do not sense the relation between these things. You can't tell the difference between a fox ripping open a hare, and a fox accidentally ripping itself open on a sharp stick and then a hare wandering past its corpse an hour later: you can sense different intensities of odour, but the intensity depends on a raft of inscrutable variables: distance; wind currents; chemical processes affecting the odiferous chemicals (which will vary depending on the chemical, the temperature, the atmospheric composition, the amount of sunlight, etc); the strength of the initial odour; and the length of time since the smell was 'caused'; and the length of time for which it was caused (eg if a cat smell is stronger than a dog smell, was the cat there more recently, or was the cat just there for longer?).

With sight, we see events. It can then take a little brainpower to deduce the identity of the participants. With smell, you immediately smell the participants, but it can take a lot of brainpower to deduce some theory about what events they might have been involved in!

With sight, therefore, it makes sense to distinguish between what you saw happen, and what you merely deduced must have happened. But if you rely on smell, it makes little sense to distinguish between what you smelled happen and what you merely deduced must have happened, since everything is the latter and nothing is the former.

Contrariwise, an olfactory species would put great evidential weight on testimony regarding existance and presence. You might, for example, see evidentiality embedded not in verbs (which prototypically relate to events), but in nouns themselves (which prototypically relate to objects). So, in Human, it makes sense to distinguish "the man ate-1V the duck" (I saw that [the man ate the duck]) , "the man ate-1D the duck" (I reckon that [the man ate the duck]) and "the man ate-3 the duck" (I've heard that [the man ate the duck]). But in Dog, this wouldn't make as much sense. Instead, you might distinguish "the man-1V ate the duck" (the man (whose existence I directly smelled) ate the duck), "the man-1D ate the duck" (the man (whose existence I deduce) ate the duck) and "the man-3 ate the duck" (the man (about whom I've heard) ate the duck).




[echolocation is very similar to sight - it's constructive, probably even more so. Likewise electro- and magnetoreception. Taste is more like smell - it's recognitive, though less so (there are fewer tastes). Touch is in this respect in theory constructive. The problem with taste and touch is that they're different in another way: they're contact senses, and hence they only report the event of contact itself, not events in the world. In this respect, touch is in practice more similar to recognitive senses, unless it's extended in some extreme way (eg for a spider in a web maybe it's more similar to a constructive sense, or if you're a giant fungus with a vast web of microtendrils throughout a huge area of soil....). Sound is actually the weirdest sense in this schema, as it combines constructive and recognitive elements: our senses of amplitude, rhythm, and direction are constructive (we can hear a collision in a certain place (an event), and reconstruct what could have collided), but our senses of pitch and timbre are recognitive (we can hear that an erhu is present, even if we aren't sure where it is exactly or what is playing it). In theory, these two types of hearing can work together to create a complex sense; but in practice, they're usually used for entirely different purposes. Our recognitive hearing is limited in what it can detect, but it's extremely acute - we can hear a distinctive ringtone, or the voice of a friend, from rooms away, even if we're distracted, and even if we can't work out where it's coming from. Our constructive hearing, on the other hand, requires active attention, and is not very acute at all (if we don't recognise a pitch/timbre pattern, we can actually completely fail to hear sounds significantly louder than that voice or ringtone!).]
Khemehekis
runic
runic
Posts: 2639
Joined: 14 Aug 2010 09:36
Location: California über alles

Re: Evidentiality in non-humans

Post by Khemehekis »

Fascinating post, Salmoneus! As an anosmic person, it was interesting to read more about how smell works for most humans.
♂♥♂♀

Squirrels chase koi . . . chase squirrels

My Kankonian-English dictionary: 70,000 words and counting

31,416: The number of the conlanging beast!
User avatar
LinguistCat
sinic
sinic
Posts: 238
Joined: 06 May 2017 07:48

Re: Evidentiality in non-humans

Post by LinguistCat »

You're right, I hadn't considered most of the things you pointed out Sal. Although I have heard that animals that are more scent oriented can often tell if a scent is older or newer, at least for things they're more familiar with. For example, a dog knowing if another dog marked a tree more recently or a long time ago, as well as things like if the other dog is in good or bad health, a potential mate or rival, etc. So if a dog that would be a rival has not been around for a while, the new dog will not react in the same way as if the rival has been around more recently. Of course, if a person is putting out wolf urine to keep away stray dogs and other small predators, then that would mess things up for the dog but that's an entirely different matter.

So it might be that creatures that rely on certain senses primarily, like scent or taste or other similar hypothetical senses, would not have evidentiality at all. One that use senses more like sight or sound as well as scent might use evidentiality, just like humans might. They wouldn't be likely to single out "scent evidentiality" even if scent was a primary sense for them. And if they do use it, it would be more likely to be marked on nouns instead of verbs.

Something like "You(-1V) talked-1D to the man-1V." could then mean "(I reckon that) you talked to the man (because I smell him on you)." Or something close? I figure second person subjects would tend not to be marked, at least in speech, until things like radio or phones became common. Or maybe I've confused myself.

For that matter, I expect someone would bring up that having two types of evidentiality for different things would likely simplify or get dropped; But the fact some languages have it at all seems excessive to me but it exists, so why not a dual system?
User avatar
eldin raigmore
korean
korean
Posts: 5797
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Evidentiality in non-humans

Post by eldin raigmore »

A language spoken by a culture that has something like hadiths might have the following two kinds of hearsay evidentials, possibly among others.
I’ll call them broken and unbroken.
Unbroken hearsay is when every transmitter in the chain knows who the originator, and every earlier transmitter, is or was, and regards them all as reliable.
Broken hearsay is when it is unknown how the saying got from one transmitter to the next known transmitter, or one of the transmitters doubts or doubted the reliability of an earlier transmitter.
....
I once proposed a misogynistic culture in whose language a “reliable hearsay” evidential was that every reporter in the chain was male, and “unreliable hearsay” or “frivolous gossip” was when at least one reporter in the chain was female.
It’s much easier to describe but much more unappetizing than the “Hadith-like” evidential above.
...
Two distinctions sometimes made in evidentiality in RL natlangs where the “direct sensory evidence” distinction is also made:
* is the direct sensory evidence available to both speaker and addressee, or only to the speaker?
* is it available at the time of the speech act, or was it only available earlier?

I might have fun with also having an “evidence only available to the addressee” value to the feature. Or a value or set of values that indicate whether or not the evidence is or was available to a third party.
Also; could it make sense to have a value indicating the evidence will be available, but only after the speech-act?

....

Fish whose ancestors never lived on dry land basically have only two senses in the water;
* sight
* feeling/smelling/tasting/hearing
User avatar
LinguistCat
sinic
sinic
Posts: 238
Joined: 06 May 2017 07:48

Re: Evidentiality in non-humans

Post by LinguistCat »

[:D] I don't think in my case I'd use something similar but if I did, I might use that for my cats.
eldin raigmore wrote: 04 Feb 2021 07:58 ...
Fish whose ancestors never lived on dry land basically have only two senses in the water;
* sight
* feeling/smelling/tasting/hearing
I'd think it would be at least three: sight, feeling/hearing, smelling/tasting. I don't think feeling and hearing would necessarily be conflated either, but I can see where that came from.
User avatar
eldin raigmore
korean
korean
Posts: 5797
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Evidentiality in non-humans

Post by eldin raigmore »

LinguistCat wrote: 04 Feb 2021 17:12 I'd think it would be at least three: sight, feeling/hearing, smelling/tasting. I don't think feeling and hearing would necessarily be conflated either, but I can see where that came from.
It’s only two.
The lateral line system accomplishes touch=taste=smell, hearing, and balance.
Since all chemosense depends on what’s dissolved in the water, and there’s no way for some chemicals to be vaporized, smelling basically is the same as tasting the water.
Touching and feeling something with part of their skin-surface that has tastebuds on it is pretty much the same as tasting it.
A touch-receptor and a tastebud are just about the same thing.
Hearing is about the same as feeling the water move or shake.

Sensory nerves in any vertebrate sense organ are about the same as sensory nerves for any other sense organ, or in any other vertebrate.
The differences between the senses depends on the rest of the sense organ, not on the nerve.

Since the fishes’ taste and smell and hearing and touch are all accomplished by the same sense organ — viz. the lateral line system — it’s extremely likely they’re all the same sense, and that’s the default hypothesis.

[citation needed] for everything, probably. I know I saw it someplace reputable; I couldn’t tell you where.
And maybe I’m wrong, and maybe you’re right.
Whatever. Cats aren’t fish, and they live in air not water, so you won’t have that problem!
Khemehekis
runic
runic
Posts: 2639
Joined: 14 Aug 2010 09:36
Location: California über alles

Re: Evidentiality in non-humans

Post by Khemehekis »

What would it be like for cetaceans?
♂♥♂♀

Squirrels chase koi . . . chase squirrels

My Kankonian-English dictionary: 70,000 words and counting

31,416: The number of the conlanging beast!
User avatar
LinguistCat
sinic
sinic
Posts: 238
Joined: 06 May 2017 07:48

Re: Evidentiality in non-humans

Post by LinguistCat »

...
Whatever. Cats aren’t fish, and they live in air not water, so you won’t have that problem!
Gratefully, yes. And while still different than humans, cats have much closer sensory inputs than between us and fish.
Salmoneus
MVP
MVP
Posts: 2232
Joined: 19 Sep 2011 19:37

Re: Evidentiality in non-humans

Post by Salmoneus »

eldin raigmore wrote: 04 Feb 2021 20:57
LinguistCat wrote: 04 Feb 2021 17:12 I'd think it would be at least three: sight, feeling/hearing, smelling/tasting. I don't think feeling and hearing would necessarily be conflated either, but I can see where that came from.
It’s only two.
The lateral line system accomplishes touch=taste=smell, hearing, and balance.
Since all chemosense depends on what’s dissolved in the water, and there’s no way for some chemicals to be vaporized, smelling basically is the same as tasting the water.
Touching and feeling something with part of their skin-surface that has tastebuds on it is pretty much the same as tasting it.
A touch-receptor and a tastebud are just about the same thing.
Hearing is about the same as feeling the water move or shake.

Sensory nerves in any vertebrate sense organ are about the same as sensory nerves for any other sense organ, or in any other vertebrate.
The differences between the senses depends on the rest of the sense organ, not on the nerve.

Since the fishes’ taste and smell and hearing and touch are all accomplished by the same sense organ — viz. the lateral line system — it’s extremely likely they’re all the same sense, and that’s the default hypothesis.
I don't think this really holds up.


Clearly, taste and touch are not "just about the same thing". They have almost nothing in common.
While hearing is on a conceptual level a form of mechanoreception - feeling the air shake on land, or feeling the water shake in the sea - it's not generally considered the same sense as touch.
So far as I'm aware, the lateral line plays no particular role in fish olfaction.

In fact, fish have two chemoreceptive senses, just like us, and these are known as smell and taste. Smell, in fish as in mammals, detects a wide variety of chemicals in low concentrations, and hence operates at a distance; smell is concentrated in the olfactory bulb. It is used to learn about the surrounding world in general. Taste, on the other hand, detects only four (or maybe a few more) types of chemical stimulus, and is very acute, but only works on physical contact, and is specialised to discern edibility. Unlike mammals, fish tastebuds aren't on the tongue - they're on the lips and face, some species have a special tasting organ in the mouth, some species have tastebuds on protruding barbels, and many have some tastebuds spread over their body, though at a lower density.

[most terrestrial vertebrates actually have three chemical senses, as they also have a vomeronasal organ. This sense, like taste, detects certain low-volatility chemicals, but isn't connected to edibility per se and detects a broader range of chemicals than taste can, so is in some ways more like smell. However, its routes to the brain are different from both smell and taste, and it's closely connected to sexual and beligerant arousal]

Fish also have three mechanoreceptive senses, unlike us. One of them is hearing, which, unsurprisingly, occurs in the ears (many fish have a clever system in which they actually hear through their swim bladders, to amplify sounds, but the sound is still ultimately detected in the ear - it's possible these fish may also 'hear' the surrounding water pressure in some sense). The second is the lateral line system. This is quite different from hearing in two ways. First, two entirely different sense organs are used. And second, they detect different things: the lateral line generally only detects low-frequency pressure waves, and is only useful at short ranges, but is extremely sensitive, and can be used to detect nearby motion, and for some fish even nearby geography (through a sonar-like system) - notably, the lateral line is a highly directional sense organ (it detects not only pressure changes but the direction of waves); hearing, on the other hand, only detects high-frequency pressure waves, and is less sensitive, and is not inherently direction-sensitive, but operates at much greater distances. And then there's the third, which we share: touch. This is detected all over the body, but has a very low sensitivity, and is only really useful for detecting direct contact.

So fish are generally said to have six (external) senses, rather than our five. However, some species have seven or more, because they also detect electric fields, and/or the earth's magnetic field. And some may also detect water pressure independently of their other senses.
User avatar
eldin raigmore
korean
korean
Posts: 5797
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Evidentiality in non-humans

Post by eldin raigmore »

Khemehekis wrote: 04 Feb 2021 22:09 What would it be like for cetaceans?
Cetaceans are descended from land-dwellers. They breathe air so for them smell and taste are different senses. Also for them hearing is a different sense.
Basically “touch” is where we leave every sense that isn’t sight or hearing or scent or taste.
Salmoneus
MVP
MVP
Posts: 2232
Joined: 19 Sep 2011 19:37

Re: Evidentiality in non-humans

Post by Salmoneus »

The question of chemoreception in cetaceans is still controversial.

Biologically, there's a lot of evidence suggesting that they shouldn't be able to taste or smell: the nasal cavities are massively reduced, the nerves that convey olfactory data are missing and some of those for taste are repurposed in many species. Many specimens completely lack taste buds, and in many species even most or all of the genes that normally create taste receptors are missing or inactive - bottlenose dolphins only have the gene for salt, and even that may not actually work. [interestingly, in toothed whales much of the chemoreceptive apparatus has been repurposed for echolocative functions]

However, the evidence is mixed and conflicting: some specimens of some species appear to have evidence of having featues associated with taste or smell (or vomerolfaction).

More importantly, behavioural evidence suggests that at least some cetaceans must have SOME chemoreception. Bottlenoses appear to be able to detect some bitterness, as well as salt - experiments on whether they can detect sweetness are so far contradictory. Other species have shown some sensitivity to blood and excretions. So far as I can see, it's not known exactly how they sense these things.

However, it does seem that cetaceans in general, particularly toothed cetaceans, have extremely reduced chemoreception compared to fish or land mammals, and it's likely that at least some species lack the distinct senses of taste and smell that humans and fish have.

-----

Some cetaceans may also have magnetoreception. We believe this mostly for biological reasons: they have internal magnetite deposits that clearly appear to be directly connected to the nervous system. Magnetoreception has also been used to explain whale behaviour - their long-distance navigation, and particularly their sometimes-baffling failures of navigation, but this is just guesswork. Frustratingly, dolphins in captivity refuse to display any sign of being able to detect magnetic fields. It may be that we haven't thought of the right tests, or that some species have it and some don't, or that none of them do.

At least one type of dolphin has electroreception. Some other species look like they could have it, but it's not been shown whether they actually do or not.

---------------

['touch' usually refers to the contact pressure sense. In a very loose sense you might also include the senses of hot and cold, and of peripheral pain. Other senses, however, are generally not considered forms of 'touch' - the human senses of proprioception and equilibrium, and animal senses like water pressure (where independent from contact pressure) and magnetic alignment]
Khemehekis
runic
runic
Posts: 2639
Joined: 14 Aug 2010 09:36
Location: California über alles

Re: Evidentiality in non-humans

Post by Khemehekis »

Fascinating stuff -- so there's a lot we still don't know. This could give a conworlder writing sapient cetaceans license to Make Shit Up.
♂♥♂♀

Squirrels chase koi . . . chase squirrels

My Kankonian-English dictionary: 70,000 words and counting

31,416: The number of the conlanging beast!
Post Reply