# Pheromone-based language

On Earth, there are many species that communicate using pheromones - from the simple (cats marking their territory) to the complex (ants conveying information about food, danger, and the current status of the colony).

In the interest of creating a highly alien-feeling species, I would like to consider the possibility of a sapient species that uses pheromones as its primary means of communication.

Pheromone-based communication has some advantages over sound (like being able to leave messages in a location), but also some disadvantages. One big disadvantage is that while sounds can be chained together indefinitely to create new concepts (phonemes combining to make words), pheromones are seemingly stuck with whatever your body is physically capable of producing and are harder to give 'structure' to, so it is hard to figure out how such a species would be able to formulate and communicate novel concepts.

How might a pheromone-based species create 'words'? How would they 'write' novel concepts? How would this system evolve? And what interesting effects would this have on their psychology and society? The more alien the answer, the better.

• Pheromones are simple hard-wired things. Ask for scent in general, where different smells can be learned and are in concious awareness. – JDługosz Jun 30 '16 at 7:19
• If an intelligent civilization arises from a species which uses scent for communication, as they probably don't have a good enough hearing, they will most likely develop a visual form of communication. They have to develop some kind of a writing system, to have any chance of a technological civilization, so they could also develop something like a sign language to communicate in rain, or against the wind, etc. – vsz Jun 30 '16 at 17:05
• Look up The Runelords by David Farland. The series contains a species that communicates this way, with extensive discussion of how it works, how they can both "speak" and "write" with scents, and the pros and cons compared to sonic communication. – Mason Wheeler Jun 30 '16 at 17:34
• @Hohmanfan's answer perfectly describes the limitations. Here's an idea: Just like facial expressions augment the human language, have your aliens augment whatever the language they are using with pheromones. Just like a young girls reddening cheeks betraying her emotions is beautiful, you'll be able to create beautiful moments in your story, if you're writing one. – sampathsris Jul 1 '16 at 11:16
• Current research indicates humans might distinguish a trillion different scents. How many could a species focused on smells distinguish or generate? Also, some research indicates that dogs might map 3D areas by smells (looking for citations). It might be far easier than we imagine, though problems remain. – user2338816 Jul 1 '16 at 18:49

One big disadvantage is that while sounds can be chained together indefinitely […] (pheromones) are harder to give 'structure'

That is an important thing to notice. Can a scent language carry enough information to be complex?

Let us first see if communication by scent really has this potential to carry a significant amount of information. As you say, sounds can be stringed together, but scents only piled. The information carried by sounds is therefore theoretically (in bits):

$$log_2(n^l)$$

Where $n$ is the number of states ("sounds") and $l$ the length of the string. In contrast, the theoretical information content of a scent pile (in bits) is simply:

$$n$$

Where $n$ is the number of possible scents. The only thing we can vary is whether a scent is in the pile or not, the order is irrelevant. Ignoring punctuation, capitalization and special characters, the alphabet encodes approximately 5 bits per sound. But as things like "fjafjkldskf7jkfdj" does not make sense, this paper says the "real" information content of English text is about 1 bit per character. Thus, the information a human get from a maxed out tweet is about 140 bits. As the information in bits carried by a scent pile is $log_2(2^n)$, simplified to just $n$, a Twitter capable species must have 140 distinct scents. But wait a minute! Such a pile can contain over a hundred different scents! What if there is some limit on how many scents you can combine before it becomes confusing? Let us have such a limit $m$. The information in bits is then:

$$log_2\left(\sum\limits_{i=0}^m\frac{n!}{i!}\right)$$

(basically, this excludes all combinations including too many scents) Of course this model does not completely model the real limitations of the information content of a language, but it is useful for comparison. At this point, it seems like it is perhaps a little bit limited for long messages, but not that bad. What would really matter is how fast the the organism can refresh its senses after registering a pile. If that does not long, it can simply receive another touch from another individual, or take a sniff from the next scented paper. In that case, the order of the scent piles is important, so the information carried is:

$$log_2\left(\left(\sum\limits_{i=0}^m\frac{n!}{i!}\right)^l\right)$$

Where $l$ is the number of scent piles. That has an even better computational class than sound strings! (But do not only look at that. For small messages, having a large amount of scents to choose from is most important, for longer messages though, the refreshing rate is what matters). Given the relatively high information content of a pile, a system with ideas or words associated with a specific pile comes to mind, much like the Chinese writing system.
(TLDR; If you do not care about the math, the answer is that yes, scent speak can carry enough information to be complex.)

The durability of a message is another consideration. It can be used for communication much like speech, with perhaps some limitations regarding the noise of large crowds. Another interesting capability is to serve the same role as sticky notes, temporarily small messages left in a place. Long term information storage is more difficult though, perhaps some scratch'n-sniff system is possible.

• Could you leave a little more information about the math? I can see your arguments and the formular you base on that. But I'm not aware of the meaning this has for the OP. – Zaibis Jun 30 '16 at 13:30
• @Zaibis Sure, I can try to make my point a little more clear, and try to make it more understandable. – Hohmannfan Jun 30 '16 at 13:35
• Isn't there a problem with scents that they linger for a long time, making subsequent communication much harder? Also, it's much harder (maybe impossible?) to determine direction. Also, speed. I can shout loudly and it can be heard in hundreds of meters, but scent takes quite some time, or maybe doesn't even arrive, if the winds are unfavorable. – vsz Jun 30 '16 at 14:53
• @vsz Yep, lingering is the problem with the refreshing rate, causing it to be favourable to have a broad range of scents for shorter messages. I think the range problem is somewhat compensated by the fact that the scent will stay in the area much longer than either sound or light. More time but less space footprint. – Hohmannfan Jun 30 '16 at 15:11
• @TemporalWolf You are correct 2^n is the number of combinations, however, I am measuring the information (measured in bits). That is log_2 of the number of combinations. – Hohmannfan Jun 30 '16 at 22:53

A species who's solo ability of communication is through scent/pheremones, would probably evolve a complex scent system as needed. Still, language would develop faster than biology, so that would only hold up for the very primitive peoples.
Base concepts through scent are easy enough, which is why it works for creatures like ants and termites, but even bees dance to convey location, because it's impossible to do through mere scent. Ants have to leave a scent trail to and from places, they can't tell each other where to go.

I'm not a linguist, but I think an important part of language is the ability to explain advanced concepts. Pheromones alone wouldn't be able to do that, so at the very least you'd have to complement it with artificial scents. But that would mean every individual needs to carry supplementary scent objects with them, just to 'speak'.

It's hard to imagine a language based solely on scent. Even a deaf-blind species is likely to combine it with a more concrete sense, like touch, in my opinion.
A language might be spoken through scent in general terms, but for specifics you'd need to touch each other in some way.

Likewise, I imagine a written version of the language would incorporate a kind of braille. A combination of feeling shapes and scents could convey more complex information than feeling and scent alone.

I think the important question here is, why would the species not use another method of supplementing communication, beyond scent?

• "why would the species not use another method of supplementing communication, beyond scent?" One possible idea is that there are multiple communicating species who can understand other species communication enough to be a risk. So this species evolved pheromone-based communication as a security measure (the other species cannot comprehend it). – called2voyage Jun 30 '16 at 16:43

Here's a new line of thought.

Pheromones, meaning strictly smells produced by the alien's body, are limiting because presumably an organism can only produce so many chemicals and certainly cannot develop new smells quickly (i.e. invent new words).

However, if this species could store samples of smelly chemicals from the environment for long periods then this would enable language in a couple ways.

First, there are an astoundingly large number of chemical molecules that a sensitive enough 'nose' (i.e. chemical receptor) could differentiate between. Imagine molecules = words and you have a large language even assuming most molecules would be poisonous to this species.

Second, there is a plausible trajectory for the development of such a language. It might start with food sharing or introducing others to an unfamiliar food. Then it evolves into a simple referential system where everything is a concrete noun. (i.e. I present you the smell of a predator to warn you. Interestingly, such a smell would be incredibly useful but incredibly dangerous to acquire). Eventually it becomes a symbolic language.

This communication method has a couple appealing 'alien' points: it could require a physiological 'smell pouch' for storing and concentrating smells. Also, it means new words can only be discovered, not invented. To me this feels like an idea that could spawn many interesting stories.

Jack Cohen's "Zarathustrians" communicated within their octets by their excreta. As we use our mouth for both speaking and eating, they use their anus for two purposes.

The concept there is that within a closed group, chemical messaging can be effective as a way of collaborative thinking. Their bodies would produce the signalling compounds rapidly and they would be mixed with those others in way that would be more like joint editing of a google doc than direct speaking as we experience.

It works better as a adjunct to other methods of communication (sound, light) or more alien (electrical or magnetic fields, unfocused but fourier analysed light - light as hearing)

One thought on how this would work in practice comes from ants.

Ants physically touch things with their antenna in order to "smell" the chemical signals present, this gives them a much greater ability to smell non-volatile chemicals.

So your scent species should have chemical sensors on their hands or other mobile antennas and would communicate mostly by being in direct contact with one another.

I wish to point another drawback of such system

## Information speed transmitting.

Let's compare crowd's alien scent capable and sound-voice capable.

These both crowd's sitting somewhere in safe places, and in each group speaker will talk to group.

Smell is carried by molecules, and how fast listener(I'll use that therm for both groups) will get information depends how fast smell diffuse in air. There are formulas for that, but as we are smell capable creatures(most of us) probably you had personal experience multiple times of how fast smell is distributed, it's not fast, compared to sound.

• actually very important moment. Any danger signal can be heard almost instant. Only smell is almost useless in fast reactions to danger as group.
In fast reaction, but overall it works, specially if you do not care about individuals.

Reach distance of smell communication is more distant then sound. There are Whistled language's or short video about with examples here (very interesting and useful concept for open spaces, and maybe not only). Optimistically, maximum distance of communications will be something like under 1km(depends, and no, I'm not sure). Butterflys(at least some) attracting each other from greater distances in time of mating.

Although there сomes difference between sound and odor. Despite what we say, which combination of sounds we use in our sentence - it will arrive how it was send (in same order of sounds). By odors that is not quite true, because their speed depends on mass of molecules, longer distance is, more significant is that effect, obliviously, it may change order of odor sounds, at some distances, and that means not $\frac {n!}{i!}$ because odors have to be emitted in order of their molecular masses. I mean, I assume they have, but really it's too complex even at level I may think about. You may be interested in reading Gas chromatography.
Sounds corruption is also a problem, but they tended to corrupt as whole word or sentence, that way most of times, we know it's broken when that happens.

For simple, actually 1-2 bit signals, odor works well. But complex sentence over 1km not sure about that, but even if it will, it will take hours instead 3sec for sound creatures.

Back to speaker's
When odor speaker speaks, he can't speak fast and too long, just as so easy as sound-speaker do. For smell it needs some time to disappear, it's like speak in volume where reverberation works in that way so it needs minutes for one word to disappear.

• there are some solutions, like media, rope-band like magnetic tape, which speacker records, and all read(circle, where speaker is only one erasing and recording point, all other just read). But that means no jokes are allowed, no speaker interruption - could be solved at some extend too - as bands of records on same tape - it may be even very useful, objection or complement may be placed exactly near that part of speech you are objecting or complementing. (we do that too just amount of bands are equal to count of peoples)

• or speaking in stream of air(from speaker to audience), then speed of air will be limiting factor of speed of information transmitting. But we are just limited by how fast we can generate sounds.

• they will have hard time to whispering information which is important just between two(spies)

## Speed

Just speed of transmitting of odors, it not actually speed which I wish to point.

In voice communication One human may easily (if other agree) to send message fast and easy to 100 people - just stand and speak.

Although by odor communication you may inform millions at a time, not fast not easy, but still that millions.

For our human there will be at least 3 steps in communications(rumors) to spread information over million peoples, and each step may contain errors (misunderstood, lie, what ever), but even that in form of rumors it quite fast. Although with internet, informing millions of people, who wish to be informed, is quite easy task: fast, reliable, exact (big deal actually) (yes yes, I know, not so easy, but you have to agree there is Big difference)

Actually rumors, is good enough model, to imagine how such odor communication may look like, in therms of information transmitting. It's possible to get some true bits over some time, but it's not precise, and takes time and effort.

## Conclusion

In short therm odors may have advantage, specially for simple communication systems, for long and short distances. And it works well for simple creatures, as beacons (shark ability to smell) etc. But it have serious disadvantages for more complex communication systems.

Can not say, that it's impossible to make it work for complex systems too, but making it realistic definitely not an easy task.

Take best of two worlds(sound or tactile or both), it's good for relatively longer lasting information(minutes, hours), as emotion expressions as example or maybe other tuning factor for information transmitted, sticky notes you mention. Handwave a bit. Really, in this case, I do not wish to be like those humans who told rockets are impossible, but it's long way to figuring out, what is good for, and limitations of it.

We had and have different approaches to transmit information: sound, visual, smell, artifacts, flowers, combination of symbols(of anything, really anything), Quipu. You may wish to start by investigating how has been solved such problems in the past, and how we do that today. We exploiting all our sensing abilities, not in equal proportions, not always, and it's just so happened we are not the best at smell, but even then we use it as information transmitting and information source, obivious example Perfume. I smell food each time, because it tells a story about itself, so at least I listen to my food each time, before we join together))

Excreting and sending pheromones over a distance takes time whereas speech is quite fast. This doesn't make it impossible. It is possible for your species to have more "sounds" than a human can produce and differentiate. You could use each smell as a syllable or even a word so that speed of communication would be increased. Additionally, to reduce confusion during communication, your species could release pheromones that decay quite fast. This will allow them to talk even faster (still much slower than speech). But it will take away their ability to leave messages.

I believe these species could reach sentience, using their communication ability for their advantage and having bigger brains to improve their chances of survival. If they can excrete semi permanent smells, they could map out regions for possible dangers and food sources. But I think it will be hard for them have a scientific community due to teaching to others would take a lot of time.

One interesting aspect of their civilization would be the literature in their society. Image them building roads to "write" poems on them. Walking on these roads would be very strange for humans as the smell keeps changing as they walk.

As others have pointed out, touching with antennae allows for more discrete communication than scents in the wind. It allows for a kind of "almost tactile" braille, so it can be sequential and have structure. Sequence and structure take the aliens far beyond the "one smell = one word" or "one smell = one sentence" limitation. A small set of scents would be sufficient to express an infinite number of ideas in the same way a small set of phonemes in a language combines to make hundreds of morphemes and morphemes combine to make hundreds of thousands of words.

As an example,consider three "scent piles" deposited and "smelled" in sequence:

((cinnamon-basil-lavender) (rose-basil-rotten eggs-lemon) (pepper-onion-ginger))

Using the scent piles as phonemes (scentemes?), the alien species could combine them into words. If the scents are deposited in a pattern, the spatial layout can also carry information like braille or the sub-parts of Chinese characters.

Using the scent piles or combinations of them as words deposited in sequence, the alien species could take advantage of syntax.

If the order of the scent piles in the example above was "Dog bites man"(or "Alien bites man", as the case may be), then the same scent piles in the reverse syntactic order would be:

((pepper-onion-ginger) (rose-basil-rotten eggs-lemon) (cinnamon-basil-lavender))

meaning "Man bites dog". This simple sequence of scent piles can carry an underlying structure which can be represented as a tree structure that, in turn, can express more complex ideas. With syntax, the combinations are infinite.

Granted, the scent from the first scent pile may drift or bleed to the next, clouding the sequence, but phoneme sequences suffer from this kind of clouding too, with the first phoneme coloring the next. Languages take this in stride, so the aliens could still express complex ideas as long as the bleeding doesn't completely obliterate the sequence.

If the aliens had antennae that smell or taste, taking advantage of "scent spots" over "scent piles" would bring them closer to using a kind of braille.

One could imagine aliens with skin pores that could secrete scent piles or scent spots in a pattern on the skin, so it wouldn't be limited to depositing on surfaces in the environment.