Formal Monkey Linguistics

This page answers questions about four recent papers in primate linguistics:

• • Schlenker, Philippe; Chemla, Emmanuel; Schel, Anne; Fuller, James; Gautier, Jean-Pierre; Kuhn, Jeremy; Veselinović, Dunja; Arnold, Kate; Cäsar, Cristiane; Keenan, Sumir; Lemasson, Alban; Ouattara, Karim; Ryder, Robin; Zuberbühler, Klaus: 2016, Formal Monkey Linguistics. Target article in Theoretical Linguistics.

  • Preprint version Published version

• Schlenker, Philippe; Chemla, Emmanuel; Schel, Anne; Fuller, James; Gautier, Jean-Pierre; Kuhn, Jeremy; Veselinović;, Dunja; Arnold, Kate; Cäsar, Cristiane; Keenan, Sumir; Lemasson, Alban; Ouattara, Karim; Ryder, Robin; Zuberbühler, Klaus: 2016, Formal Monkey Linguistics: the Debate. (Replies to commentaries). Theoretical Linguistics.

  • Preprint version Published version

  • Schlenker, Philippe; Chemla, Emmanuel; Arnold, Kate; Zuberbühler, Klaus: 2016, Pyow-Hack Revisited: Two Analyses of Putty-nosed Monkey Alarm Calls. Lingua [the editorial team founded Glossa after our paper was accepted for publication].

  • Preprint version Published version

  • Schlenker, Philippe; Chemla, Emmanuel; Cäsar, Cristiane; Robin Ryder; Zuberbühler, Klaus: 2016, Titi Semantics: Context and Meaning in Titi Monkey Call Sequences. Natural Language & Linguistic Theory.

  • Preprint version Published version

It also makes reference to

• • Schlenker, Philippe; Chemla, Emmanuel; Arnold, Kate; Lemasson, Alban, Ouattara, Karim; Keenan, Sumir; Stephan, Claudia; Ryder, Robin; Zuberbühler, Klaus: 2014. Monkey Semantics: Two 'Dialects' of Campbell's Monkey Alarm Calls. Linguistics & Philosophy 37(6): 439-501. DOI 10.1007/s10988-014-9155-7

  • Preprint version Published version

discussed more fully on a separate page.

The main results are also summarized in a recent 3 1/2-page manuscript; an expanded version appeared in Trends in Cognitive Sciences:

Penultimate version (final version to appear in TICS): [pdf]

Published version [gated]

An encyclopedia article will summarize our main results: [LingBuzz]

A popular interview in French was given by Alban Lemasson and Emmanuel Chemla on some issues raised below and in our 2014 paper: [video]

A later survey paper on general calls by Dezecache and Berthet is also relevant: [gated] [poster]

A related paper on evolutionary models of call meaning:

Steinert-Threlkeld, Shane; Schlenker, Philippe; Chemla, Emmanuel: to appear, Referential and General Calls in Primate Semantics. Linguistics & Philosophy.

Note: Some of the same team members also published:

• Maarten Versteegh, Jeremy Kuhn, Gabriel Synnaeve, Lucie Ravaux, Emmanuel Chemla, Cristiane Cäsar, James Fuller, Derek Murphy, Anne Schel and Ewan Dunbar: 2016, Classification and automatic transcription of primate calls, Journal of the Acoustical Society of America (Vol.140, No.1).

• Preprint Published version

Small erratum: the postal code of ENS incorrectly appears in the paper as 75011. It should be 75005.

Q&A on 'Formal Monkey Linguistics' (and 4 related articles, all published or in press)

Note prepared by Philippe Schlenker, with Jeremy Kuhn's help (last update: June 21, 2016)

Quick links to questions:

• • What do these articles try to achieve?

  • Which monkeys are discussed in these articles, and how are they related to humans?

  • So what do monkeys talk about?

  • What are the main 'take home' messages of these studies?

  • What are your main empirical findings on these call systems?

  • What are the main theoretical findings?

  • So does this mean that monkey languages(*) can be as sophisticated as human languages?

  • What do we learn about the evolution of language from these call systems?

  • Why did you start primate linguistics with monkeys rather than apes, which are more closely related to humans?

  • Your articles always discuss at least two theories, and sometimes more. Is this all sheer speculation?

  • What are the next steps?

  • What are the long-term perspectives?

  • Why are your articles published in linguistics journals?

  • And finally... what are the worst misconceptions that your work has given rise to?

What do these articles try to achieve?

Like 'Monkey Semantics' (published in December 2014), these articles apply methods from contemporary linguistics to the analysis of monkey calls. They thus seek to contribute to an emerging field of 'primate linguistics'.

Which monkeys are discussed in these articles, and how are they related to humans?

These articles discuss the call systems of some African and South American monkeys. In some estimates, African monkeys share a most recent common ancestor with humans approximately 30 million years ago. As for American monkeys, one might have to go back in time more than 40 million years to find a common ancestor with humans. So these are not our closest cousins! (For comparison, we share a most recent common ancestor with chimpanzees – as well as with bonobos – approximately 6 million years ago.)

In greater detail: we studied two subgroups of Old World monkeys, namely Cercopithecines (Campbell's monkeys, Putty-nosed monkeys, Blue monkeys) and Colobinae (Black-and-White Colobus monkeys, King Colobus monkeys). In addition, we studied one species of New World monkeys, Titi monkeys.

So what do monkeys talk about?

The calls under study are primarily alert calls that are related to dangers in general and predators in particular. One shouldn't infer that there are no other calls. For instance, female Campbell's calls have rich inventories of social calls, but these are hard to study than male alert calls (among others because there are several females per group and because their calls are less loud) – which is the reason much research is concerned with alert calls.

What are the main 'take home' messages of these studies?

• • The main message is that rich data collected by primatologists are ripe for linguistic analysis, and that methods from contemporary linguistics can illuminate them.

  • From this it doesn't follow that monkey languages and human language share non-trivial properties (although the general methods of formal linguistics can illuminate both).

  • Still, one key tool of contemporary linguistics, that of 'implicatures' (explained below), plays a key role in our analyses.

  • Finally, a by-product of our 'comparative monkey linguistics' is that aspects of the evolution of monkey calls can probably be reconstructed over millions of years (this was known in the 1970's, but this kind of work fell out of fashion). We thus call for the investigation of 'evolutionary monkey linguistics'.

What are your main empirical findings on these call systems?

Each species studied conveys information by way of interesting systems of alert calls.

(i) As we had discussed in our 2014 paper, 'Monkey Semantics' (which in turn built on earlier work by primatologists Ouattara and colleagues), Campbell's calls have a distinction between roots (especially hok and krak) and suffixes (-oo), and their combination allows the monkeys to describe both the nature of a threat and its degree of danger. For instance, hok warns of serious aerial threats – usually eagles – whereas hok-oo can be used for a variety of general aerial disturbances; in effect the suffix -oo serves as a kind of attenuator.

(ii) Putty-nosed monkeys have two main calls of interest, pyow and hack, in addition to booms (the latter are also present in Campbell's monkeys, a point to which we return below). In most situations, pyows is used non-specifically, as a general alert call, while hacks are raptor-related. But in sophisticated field experiments, primatologists have shown that a small number of pyows followed by a small number of hacks have a highly specific function, namely to trigger group movement. Hence a puzzle: is the meaning pyow-hack sequences derived from their component parts, and if so how?

(iii) Blue monkeys share a most recent common ancestor with Putty-nosed approximately 2.5 million years, and their call systems are in part quite similar: Blue monkeys seem to have counterparts of pyows and hacks, and booms, and also other calls not shared with Putty-nosed monkeys. We were interested in comparing the two call systems (and didn't find clear cases of pyow-hack sequences in Blue monkeys, although more work is needed on that point).

(iv) Black-and-White Colobus monkeys have two main calls, snort and roars. Snorts given singly are indicative of terrestrial mammals, roars alone are usually indicative of aerial predators. But interestingly snort-roar sequences appear to be used as very general alert calls – hence a puzzle (similar in form to the one we had with Putty-nosed pyow-hack sequences): is the meaning of snort-roar sequences derived from their component parts, and if so how?

(v) Titi monkeys were shown in field experiments by Cäsar et al. to display an extraordinary pattern in which they re-arrange two calls to convey information about the nature as well as the position of the predators – hence a puzzle: do Titi monkeys have complex syntactic and semantic rules that allow them to express rich information with just two calls?

An example is given below, with the A-call in orange, and the B-call in green. Each block (raptor in the canopy, raptor on the ground, cat on the ground, cat in the canopy) gives rise to stereotyped sequences, and these are distinctly different across conditions. In other words, each situation gives rise to a distinct stereotyped sequence.

What are the main theoretical findings?

The main theoretical finding is that in almost all cases two simple tools of contemporary linguistics make it possible to offer a simple analysis of the call systems under study.

–First, we provide a precise specification of the meaning of each call type (e.g. general alert for Titi B-calls, serious non-ground alert for Titi A-calls), and we assume that each call contributes its meaning independently from the others, hence without necessarily having a non-trivial syntactic structure.

–Second, we crucially make use of a rule of 'meaning enrichment' used in human languages, called 'implicatures'. It posits that the meaning of a word can be enriched when it competes with a more informative alternative – for instance, "possible" competes with "certain", which is more informative, and for this reason "possible" usually comes to mean "possible but not certain" (for instance in: "It's possible that the Democrats will win the elections" – which implies that this is not a certainty). In some cases, we use further mechanisms of meaning enrichment, for instance a rule according to which a call that provides information about the location of a threat should come before one that doesn't.

With these mechanisms, as well as detailed assumptions about the environment, we show that we do not need to posit complex complex syntactic or semantic structures to analyze most of the patterns that are currently understood in the monkeys under study.

Still, there are two interesting cases in which complex calls might be needed. One pertains to the suffix -oo in Campbell's monkeys, which we continue to analyze as a bona fide suffix (both morphologically and semantically), as we did in our 2014 study. The second case pertains to Colobus snort-roar sequences, which might have to be analyzed as complex units – although more work is needed to address this issue.

So does this mean that monkey languages(*) can be as sophisticated as human languages?

Not at all. First, in most cases we posit much simpler rules than are found in human in language – and in particular no complex rules of syntactic or semantic composition. Second, the mechanism of meaning enrichment we posit ('monkey implicatures') has an analogue in human language, but the rule is so natural (it is just an injunction to be as specific as possible) that it could have developed independently in unrelated systems (our articles show that that as simple as this rule may be, it has far-reaching consequences for the analysis of various call systems). Third, the root-suffix structure we find in Campbell's calls is very interesting in its own right, but far less complex than what is found in the morphology and syntax of human languages.

Thus we think that monkey languages are extremely interesting in their own right, and should be the object of detailed studies using the general methods of formal linguistics; but that from this it does not follow that the the two kinds of objects (monkey languages vs. human languages) are similar or connected in evolutionary history.

(*) Terminological note: as stated in our 2014 study, "we freely apply linguistic terminology to these calls [including the term 'language'], with the belief that they can and should be studied as formal languages with a sound system, a lexicon, a morphology, a syntax, a semantics, and a pragmatics. Importantly, we do not take a stand on the relation that these systems bear to human language; to say that they can be studied as formal systems does not imply that they share non-trivial properties with human language, nor that they share an evolutionary origin with it."

What do we learn about the evolution of language from these call systems?

For the reasons mentioned above, we are not in a position to make claims about the relation between these monkey languages and human language. On the other hand, a by-product of our research in comparative monkey linguistics is that monkey calls are remarkably well-preserved over millions of years. This fact was well-known in the 1970's, when acoustic similarities among calls was used to reconstruct phylogenetic trees (i.e. 'family trees' of different species), at a time when DNA methods were less readily available than they are today. We propose to turn the problem on its head and to start from established phylogenies (thanks to DNA methods) in order to study the evolution of monkey calls.

A simple example will make the program clear. We have reproduced below a recent phylogenetic tree of a large family of monkeys (cercopithecines) that includes several species we studied. They all share a most recent common ancestor approximately 10 million years ago. We wrote in blue the species that have booms, a loud, low-pitched call produced in a very particular way (with air sacs), and which is used only situations that are not predator-related (as mentioned above, among our 'subjects', Campbell's monkeys, Putty-nosed monkeys and Blue monkeys all have booms). Remarkably, there are two coherent subfamilies that have booms. Members of the top one share a most recent common ancestor approximately 2.5 million years ago. It is natural to posit that this ancestor had booms as well, and that these were preserved in all descendants. Of course it's not logically impossible to posit instead that booms developed independently in members of this subfamily, but this would seem to be less parsimonious. Similarly, members of the lower group share a most recent common ancestor more than 5 million years ago, and it is plausible that in that family as well booms have been present for millions of years. (A more precise assessment would require methods that go beyond our research; our goal in that part of our work was just to highlight the interest of evolutionary questions studies of monkey calls.)

The case of booms is particularly clear but not isolated – we mentioned above that Putty-nosed monkeys and Blue monkeys, which have a most recent common ancestor approximately 2.5 million years ago, plausibly share at least three calls in their repertoire. Thus we hope that future research will lay the groundwork of an 'evolutionary monkey linguistics'.

Why did you start primate linguistics with monkeys rather than apes, which are more closely related to humans?

As things stand, ape calls are less well understood than (some) monkey calls. One possible reason is that ape communication is generally multi-modal, with gestures playing an important role (monkeys that don't see each other must rely more exclusively on calls). It might also be that the content of ape calls is more complex (and less directly predator-related), although we don't know this for a fact. In any event, to establish the beginning of a primate linguistics, it seemed wise to start from the clearest cases – hence monkey rather than ape calls.

Your articles always discuss at least two theories, and sometimes more. Is this all sheer speculation?

Our research is decidedly theoretical, in that it seeks to offer detailed formal analyses of all the data available. Quite generally in science, several theories are compatible with the available data, although (i) they may incur different costs (e.g. one theory might be much simpler than a competing one), and (ii) they often make different predictions about further experiments that could be undertaken. The case is no different here, but as theoreticians we are very careful to lay out diverse theoretical options (although we usually express a clear preference for one). In addition, we are particularly cautious in our conclusions because we are at a very early stage of primate linguistics, and various aspects of monkey calls (e.g. their syntax, their use, and also aspects of their categorization ) are ill-understood. Data from future observations and field experiments will no doubt force us to modify parts of our analysis. Still, it counts as progress that very explicit theories can now be stated and compared on the basis of existing data and new predictions.

What are the next steps?

We are now hoping to make the program of a 'primate linguistics' known outside of linguistics, and to extend it to apes. We hope that primate gestures and facial expressions can also be taken into account in these linguistic analyses.

What are the long-term perspectives?

Besides contributing to the development of a primate linguistics, we hope that our methods will allow for a much more precise analysis of systems of animal communication, with much interaction between linguists and ethologists. In the long term, any theory of human language evolution should depend on a precise comparative analysis of various systems of animal communication.

Why are your articles published in linguistics journals?

Since our goal is to apply methods from formal linguistics to monkey calls, we published our papers in visible peer-reviewed linguistics journals. Our 2014 paper, 'Monkey Semantics', was published in Linguistics & Philosophy, one of the preeminent semantics journals. Our article on Putty-nosed calls, 'Pyow-Hack Revisited', is published in a generalist journal in formal linguistics, Lingua (the paper was accepted for publication by the former editorial team, which went on to found the online journal Glossa). Our article on Titi monkeys is forthcoming in another visible general linguistics journal, Natural Language & Linguistic Theory. Finally, a survey of all these results, with several additional studies, was invited as a target article in Theoretical Linguistics (with open peer commentaries). One of our goals is to convince other linguists to start collaborating with primatologists on the analysis of primate communication, and thus it is important to show that various visible linguistics journals are willing to publish work in primate linguistics.

And finally... what are the worst misconceptions that your work has given rise to?

–People love to think that monkeys have language just like humans do. We claim no such thing. In fact, the monkey languages we study have very different (and much simpler) rules than human language.

–It's sometimes said that we have found ways to 'translate' monkey utterances. That's at best a half-truth, because there is still much uncertainty about how these call systems work: we usually give arguments pro and contra various theories, and remain cautious in our theoretical conclusions.

Press releases

NYU (Linguists Team up with Primatologists to Crack the Meaning of Monkey Calls)

Note: the initial version contains a typo. “We can now study the form and meaning of monkey calls using methods from theoretical titi monkeys linguistics,” should read: “We can now study the form and meaning of monkey calls using methods from theoretical linguistics,”

CNRS/INSHS

Media

Science Daily (reprint of the NYU press release)

Note: the initial version contains a typo. “We can now study the form and meaning of monkey calls using methods from theoretical titi monkeys linguistics,” should read: “We can now study the form and meaning of monkey calls using methods from theoretical linguistics,”

Daily Mail Online

New York Magazine

GeeksOnEarth

Deccan Chronicle (India)

Financial Expression

SBS (Australia)

Futura Sciences (France)

Galileu

Le Figaro (France), August 11, 2016, Le Chat et Le Singe. Excerpt.

Futura Planète