First, I do a straightforward test of whether word order is correlated with the number of children you have.  This comes out as significant!  I wonder if  having more children hanging around affects the adaptive pressures on langauge?  However, I then show that this result is undermined by discovering that there are other linguistic variables that are even better predictors.

I used the World Values Survey:  a large database of survey results from thousands of people around the world, including what language they speak and how many children they have.  I then linked this up with linguistic typology data from the World Atlas of Language Structures.  This includes information on the basic word order of each language.

The hypothesis was that people who used particular basic word orders would have more children.  Testing this hypothesis directly, basic word order is a significant predictor of the number of children a person has (linear regression, controlling for age, sex, if the person was married, if they were employed their level of education and religion, t-value for basic word order = -18.179, p < 0.00001, model predicts 36% of the variance).  It turns out that speakers of SOV langauges have more children than speakers of SVO languages, while speakers with no dominant order have the fewest children on average (there wasn’t enough data for other word order types).  Indeed, in the strict interpretation of Evolution, SOV order is the fittest variant, since it is linked with having more offspring:

An explanation might be found in information theory:  When hearing a sentence, you want the most unpredictable information at the start.  If you only have one child, then if you know the sentence is about them (the Subject) what you want to know next is what they’ve done (the Verb).  For instance, “Harry smashed the window”.  Hence, SVO order.  However, if you’ve got more than one child, then what you want to know after the perpetrator is who they’ve done it to.  For instance “Harry Hannah Hit”.  Hence, SOV order.

Actually, a more interesting hypothesis is that having more children around influences the learning pressures on language.  If children find SOV easier to process or learn, there is a pressure on langauge to change to fit their cognitive niche.  Therefore, we should expect societies with more children to exhibit this word order.  Indeed, some previous studies suggest that SOV order is the most basic or ancestral form historically (Luke Maurits, in press, RT coverage, paper, video lecture,  Gell-Mann & Ruhlen, 2011 – see post).  However, Maurits’ work suggests that SVO is actually more efficient from an information-theoretic perspective.  Do children have different cognitive biases?  Alternatively, if children can negotiate their communication system with their parents (as Suzanne Quay argues), then they might push for SOV order more often.  I’m not sure there’s a lot of support for this, though.

Relative strength

The results above show the absolute strength of the relationship between word order and number of children.  However, given that there are many links between social and linguistic variables, and indeed, we should expect cultural variables to be correlated at a rate greater than chance due to geographic diffusion.  A much more robust approach is to hypothesise that basic word order will predict the number of children a speaker has better than any other linguistic variable.  Therefore, I tested every linguistic variable in the WALS.

I ran a number of linear regressions with the number of children as the dependent variable and the linguistic variable as the independent variable, controlling for age, sex, if the person was married, if they were employed, their level of education and their religion. We can then look at the top predictors by the magnitude of the F-score of the coefficient for the linguistic variable.  Here are the top 11 predictors out of the 142 linguistic variables:

The order of Subject, Object and Verb is the 11th best linguistic predictor of the number of children a person has. Christos was on to something.

Order of Subject, Object and Verb are in the top 15% of linguistic variables.

A stepwise regression including the control variables and the top 15 linguistic variables resulted in the following model which accounted for 37% of the variance (Linguistic variables sorted by significance at the top).  The order of Subject, Object and Verb does make it in, but it is the weakest linguistic predictor.  I’ve included the results for religions.  Note that this data encodes religions beliefs but also geographic region.

The jargon:

• Estimate = the size and direction of the relationship (in number of children). Positive values means a positive relationship. E.g. you’ll have on average 0.15 children less as your level of education increases.
• Std. Error = how well the variable fits the data
• t value = the strength of the correlation. Large positive values means large
• p = the probability that this correlation occurred by chance

Some other linguistic factors

Distance contrasts refer to deictic expressions such as ‘this’ (near speaker) and ‘that’ (further away from speaker) in English.  People with more children tend to have more specific contrasts:

This makes sense, since if there are more children running around, you’ll need more specific demonstratives to refer to them.

Another interaction was with ways of marking nominal (Bobby is a child) and locational (Bobby is in the garden) meanings.  English has only one form for these two meanings (is), but  other languages have separate forms.  People with more children tend to speak languages with separate ways of  marking the nominal and locational:

Some other interesting patterns emerged:  People who say “Two children” (Numeral before noun) have fewer children than people who say “children two” (Noun before numeral).  Below is the graph for the gender distinctions in pronouns:

But my favourite outcome is for distributive numerals. A sentence like “John and Mary have two children” has two possible interpretations:  Either John and Mary have a total of 2 children between them, or John has 2 children and Mary has 2 other children.  In English, we would distinguish these meanings lexically by putting ‘each’ or ‘between them’ at the end of the sentence.  Some languages indicate this difference with morphology.  According to the data, having no strategy for indicating this difference is linked to having more children:

Here’s my explanation:  If a person who has no way of distinguishing the two meanings of the sentence above, when they hear “John and Mary have two children”, they might think “Damn, John and Mary have 4 children!  I’d better get some more of my own … “.  This leads to a runaway affect of people having more children in order to keep up with what their neighbours.

Weak Explanatory Power

Of course, these theories are crazy.  However, their plausibility does not derive from the correlation – I could come up with even wackier stories about why these variables were connected and they would be equally supported by the correlations.  As James Winters and I argue (Roberts & Winters, 2012, see here), these kinds of statistical tests are good for generating hypotheses, but have very weak explanatory power.  They need to work together with idiographic, experimental and modelling approaches in order to support the mechanisms they suggest.

Sean Roberts, & James Winters (2012). Constructing Knowledge: Nomothetic approaches to language evolution Five Approaches to Language Evolution: Proceedings of the Workshops of the 9th International Conference on the Evolution of Language

Gell-Mann, M., & Ruhlen, M. (2011). The origin and evolution of word order Proceedings of the National Academy of Sciences, 108 (42), 17290-17295 DOI: 10.1073/pnas.1113716108

Maurits et al. (in press). Why are some word orders more common than others? A uniform information density account Advances in Neural Information Processing Systems, 23,

The Centre for Philosophy of Science of the Faculty of Science of the Portuguese University of Lisbon is organizing a 3-day international colloquium entitled ”From Grooming to Speaking: recent trends in social primatology and human ethology”, on September 10-12^th, 2012.

Conference website

http://cfcul.fc.ul.pt/linhas_investigacao/Philosophy%20of%20Life%20Sciences/int_col/index.htm

Plenary talks will be given by

Johan Bolhuis
Constança Carvalho
Augusta Gaspar
Nathalie Gontier
Mary Lee Jensvold
Simone Pika
Tim Racine
Jeroen Stevens
Jordan Zlatev

More tba

Scientific committee

- Rod Bennison, CEO Minding Animals International

- Rudie Botha, University of Stellenbosch, South Africa and University of Leiden, the Netherlands

- Daniel Dor, Tel Aviv University, Israel

- Luc Faucher, UQAM, Candada

- Nathalie Gontier, Free University of Brussels, Belgium (chair)

- David Leavens, University of Sussex, UK

- Robert Lickliter, Florida International University, US

- Jorge M.L. Marques da Silva, University of Lisbon, Portugal

- Mark Nelissen, University of Antwerp, Belgium

- Eugenia Ramirez Goicoechea, UNED, Spain

- Emanuele Serrelli, University of Milan, Italy

- Chris Sinha, Lund University, Sweden

- James Steele, University College London, UK

- Ian Tattersall, American Museum of Natural History, NY

- Natalie Uomini, University of Liverpool, UK

- Arie Verhagen, University of Leiden, the Netherlands

- Luis Vicente, University of Lisbon, Portugal

Call for abstracts

Deadline for submissions is June 30^th, 2012.

We call for primatologists, ethologists, anthropologists, sociobiologists, evolutionary, cognitive and comparative psychologists, biolinguists, evolutionary linguists, bio-ethicists, philosophers and historians of science, to provide talks on:

(1) Historical reviews on the introduction and use of primate studies to acquire knowledge on the origin and evolution of communication and language

• The rise of comparative psychology, ethology, primatology, sociobiology, evolutionary psychology, evolutionary linguistics, and evolutionary anthropology
• Cross-fostering experiments, experiments that had as goal to learn non-human primates to talk or sign, or to learn artificial languages such as Yerkes
• The shifts from behaviorism and instructionism to cognitivism and selectionism
• The nature/culture debate
• The innate/acquired debate
• The continuity/discontinuity debate

(2) Methodologies of primate communication and language research

• Which research methodologies combine and diversify ethologists, primatologists, sociobiologists, anthropologists, evolutionary psychologists and evolutionary linguists? (ASL and Yerkes experiments; instructionist, behavioral versus selectionist, adaptationist approaches; the use and disuse of Tinbergen’s 4 questions in ethology; how to study ultimate and proximate causes of behavior)
• Did classic ethology and comparative psychology, with its focus on instructionist and behaviorist methodologies, fail? Did the cognitive turn succeed in providing answers there were behaviorism failed? And is selection theory able to provide answers to questions neither ethologists nor cognitivists could?
• Which methodologies are used to study (human) primate verbal and non-verbal communication strategies in wild, captive, and natural settings (how are experiments set up, how are biases controlled, how is data collected and interpreted, how are theories formed)?
• How do ontogenetic studies of normal and pathological behavior lend insight into phylogeny (what aspects of development enable or disable scientists to draw inferences on human evolution, what’s the rationale behind comparative research, how do pathologies lend insight, either into normal development, or into the evolutionary past of hominins)?
• How do the primate and ethological research methodologies differ from, relate to, or complement genetic and neurological research?

(3) Theories on primate communication and the evolution of language

• Gestural versus vocal origin theories (grooming as gossip theories, mirror neurons, non-verbal communication theories (including facial expressions, pointing and gestural research), co-verbal gesturing theories, signing theories, mimesis, imitation).
• Evolutionary theories on language as a social communication device
• Theory of Mind versus embodiment theory, in human and non-human primates
• Theories on learning (conditioning, observational learning, imitation)
• Theories on cultural transmission (chimpanzee, bonobo and human cultures)
• Which theoretical frameworks and evolutionary mechanisms enable adequate explanations on language evolution (natural selection, drift, systems theory, the Baldwin and ratchet effect, co-evolutionary theories, dual inheritance theories)

(4) Ethical issues in social primatology and human ethology

• Policy and guidelines on (human) primate studies in the wild, under captivity, or under experimental conditions
• Animal rights (e.g. if non-human primates have ToM, do we need to attribute them legal rights, does the concept of “legal person” apply to non-human primates)
• The role and responsibility of researchers

Much more than provide a platform for the dissemination of new research results, the conference organizers will give preference to reflexive talks, that deal with theoretical, methodological and ethical issues of primate research and ethology, and how the latter fields provide insight into human language evolution.

Proceedings

A selection of talks will be published in an anthology for the Springer Book Series “Interdisciplinary Evolution Research”. Editors-in-chief of the series are Nathalie Gontier and Olga Pombo.

Submission guidelines

http://cfcul.fc.ul.pt/linhas_investigacao/Philosophy%20of%20Life%20Sciences/int_col/call.htm

Kind regards,

Prof dr Nathalie Gontier

Philosopher of evolutionary sciences
http://vub.academia.edu/NathalieGontier

Edit: A longer version of my interview at EU:Sci is now available online, Listen here!

Here’s my opinion in a nutshell: This is a great volume and I’ve really learned a lot from reading it. The authors have done a great job trying to be accessible to an interdisciplinary audience. It’s  a great place to start if you’re interested in language evolution or want to get a quick overview of a specific topic in language evolution research. I would’ve liked it if the chapters had a “Further Reading” section, however (like  Christiansen and Kirby’s 2003 volume). Some chapters felt a bit too short for me (Steven Mithen‘s chapter on “Musicality and Language” for example is only 3 pages long, Merlin Donald‘s chapter on “the Mimetic Origins of Language” is 4 pages long). I also feel that some topics, like language acquisition, could’ve been dealt with  more extensively, but then again, if you compile a handbook, it’s impossible to make everybody happy. Other recent book-length overviews of language evolution (e.g. Fitch’s 2010 book and Hurford’s 2007 and 2012 tomes) are more detailled, but also more technical and not as comprehensive and don’t cover as many topics. To quote my review:

Overall, the Oxford Handbook of Language Evolution is a landmark publication in  the field that will serve as a useful guide and reference work through the  entanglements and pitfalls of the language evolution jungle for both experienced  scholars and newcomers alike.

One last thing I’m particularly unhappy about is that the handbook doesn’t have an Acacia Tree on the cover - which seems like a missed opportunity (kidding).

I’ll try to write about some of my favourite chapters in more detail somewhere down the road/in a couple of weeks.

Jäger presented work at the workshop on Visualization of Linguistic Patterns and Uncovering Language History from Multilingual Resources at the recent EACL conference last month.  He uses the Automated Similarity Judgment Program (ASJP) database, which contains 40 words from the Swadesh-list (universal concepts) for around 5800 languages (including Klingon!).  The words are transcribed in the same coarse transcription.  The task is to calculate the distance between languages based on these lists in a way that they reflect the genetic relationships between languages.

The LDND distance measure is used as a starting point (Bakker et al. 2009):  Given two languages, the distance between them is the average normalised Levenshtein distance between each word pair.  That is, the average proportion of phonemes that need to be changed in order to transform a word in one language to a word in the other.   The measure is further normalised by the average distance between all non-synonymous words from the two languages.  This controls for the fact that if the two languages have large phoneme inventories with little overlap, the distance measure will be disproportionately high.

Weighting phoneme changes

Jäger improves this measure in two ways.  First, the distance measure is unrealistic because it treats the distance between all phonemes as the same.  Intuitively, the change between [hand] and [hent] is closer than [hand] and [mano], even though both paris differ in two phonemes.  What is needed is a way to weight each phoneme change by its linguistic plausibility.  For example, an [i] is more likely to change diachronically to a [j] than to a [k].

Luckily, there is a similar problem in bioinformatics when trying to align genetic protein sequences.  The best alignment is the one that has the most evolutionarily plausible mutations.  Phonetic representations of words can be aligned in the same way.  For instance, the Spanish and German word for ‘star’ (estrella and Stern) can be aligned as follows:

Spanish est-reya
         || ||
German  -StErn--

While this aligns two implausible changes (e and n), it also aligns a plausible one (s and S, alveolar and postalveolar fricatives).  The algorithm to discover the weights assumes that more plausible alignments will occur within related languages.  The weighting is calculated according to the following formula (based on the formula for the The Block Substitution Matrix):

where p[ij] is the probability of amino acid i aligning with amino acid j and q[i] is the frequency of i.

For phonemes, the frequencies can be calculated from the frequencies of phonemes in the ASJP database.  Jäger calculates p using the following algorithm:

1. Pick two languages from the same language family at random
2. Pick a meaning
3. Align the two words from each langauge that correspond to this meaning using a distance metric.
4. For each aligned phoneme pair, add one to that pair’s frequency

The distance metric could be the Levenshtein distance, the LDND distance or the Needleman-Wunsch algorithm (used in alignment of protein sequences).  This algorithm is repeated 100,000 times.  Because the algorithm compares synonymous words in related languages, it is more likely to sample phoneme pairs that are related by descent, and therefore similar.  This results in the following classification (from Jäger, 2012, p.4, visualised using hierarchical clustering):

The results are phonetically plausible – the vowels are together and there are sensible clusterings such as S and tS, and the nasal cluster (N, 5 n) and so on.  The frequencies can be translated into weights for use in the distance measures.

Further improvements

The second improvement on the LDND measure is to take account of the variance in as well as the average of the distance between all non-synonymous word pairs. Jäger uses an information-theoretic approach to incorporate the variance.  Essentially, this is the amount of information that you gain about a word by knowing its translation in another language (see page 4).

Jäger then compares the different similarity measurements under this approach.  Jäger’s algorithm using the Needleman-Wunsch distance metric performed better than when using the Levenshtein distance metric or the original LDND algorithm (although the effect size is small).

Visualising the differences between languages

Finally, the distances between languages were visualised using a force directed graph layout (using the free CLANS software).  Basically, each language is represented by a particle in space that is attracted to similar languages and repelled by dissimilar languages.  The CLANS program simulates the particles moving until they reach an equilibrium.

Below is the result for all languages:

And for Eurasian languages:

Jäger describes the data for Eurasian languages as following:

“The Dravidian languages (dark blue) are located at the center. AfroAsiatic (brown), Uralic (pink), Indo European (red), SinoTibetan (yellow), Hmong Mien (light orange), AustroAsiatic (orange), and TaiKadai (yellowish light green) are arranged around the center. Japanese (light blue) is located further to the periphery outside SinoTibetan. Outside IndoEuropean the families Chukotko Kamchatkan (light purple), MongolicTungusic (lighter green), Turkic (darker green) Kartvelian (dark purple) and Yukaghir (pinkish) are further towards the periphery beyond the Turkic languages. The Caucasian languages (both the North Caucasian languages such as Lezgic and the Northwest Caucasian languages such as Abkhaz) are located at the periphery somewhere between IndoEuropean and SinoTibetan. Burushaski (purple) is located near to the Afro Asiatic languages. Some of these pattern coincide with proposals about macrofamilies that have been made in the literature. For instance the relative proximity of IndoEuropean, Uralic, ChukotkoKamchatkan, MongolicTungusic, the Turkic languages, and Kartvelian is reminiscent of the hypothetical Nostratic superfamily. Other patterns, such as the consistent proximity of Japanese to Sino Tibetan, is at odds with the findings of historical linguistics and might be due to language contact. Other patterns, such as the affinity of Burushaski to the AfroAsiatic languages, appear entirely puzzling.”

The result is impressive, given that it’s a clustering of languages calculated directly from words which can be run on an ordinary laptop in a few hours.  It’s easy to imagine extensions to this by adding more cultural variables or socio-economic data.  It would be interesting to see how well the visualisation corresponded to a physical map of the world, perhaps by simulating the particles in an additional field with repelling areas representing the oceans.

Removing trees altogether

Jäger notes that the advantage of the CLANS approach is that, unlike SplitsTree or phylogenetic techniques, it does not use or estimate an underlying tree structure.  Patterns that were previously hidden may be discovered by this kind of visualisation.  However, this claim is somewhat undermined by the approach taken:  The algorithm for calculating the phoneme alignment weightings uses data about language families – which is itself part of a tree-like classification.  Without this information, the algorithm would pick unrelated words more frequently and the measure would be less accurate.  it seems that getting away from trees is difficult (see also Kevin’s post on reconstructing phylogenies)

I propose an alternative to Jäger’s algorithm which does not rely on any language classifications.  Instead of picking language pairs by family, we can pick them by physical proximity.  The languages in the ASJP database have information from the Ethnologue, including geographic co-ordinates of the center of the language community.  The alternative algorithm is thus:

1. Pick a language pair with a probability inversely proportionate to the physical distance between them.
2. Pick a meaning
3. Align the two words from each langauge that correspond to this meaning using a distance metric.
4. For each aligned phoneme pair, add one to that pair’s frequency

I actually used the cubed inverse of the great arc distance in kilometers.  It’s not clear whether this metric would work.  Clearly, languages that are physically close should have a shared history, but the measure does not take into account physical barriers like oceans or flat, open ground compared to mountainous terrain.  I tested this algorithm on Indo-European languages for 200,000 runs.  Below are the results, displayed in the same format as Jäger (2012) on the left and the same tree with IPA symbols on the right (click for larger image).

Although the results are somewhat different to that in Jäger (2012), they are also phonetically plausible.  The vowels are together, there are clusters of dental, alveolar and bilabial phonemes and there are affricate-fricative pairs.  It seems realistic, then, that this visualisation approach could work entirely without prior data on language classifications.  I tried running it on the data for all languages, but with 6 million language pairs, my poor python script was too slow!  Let me know if you’d like to see the code.

———

UPDATE:

First, between writing the paper and giving the talk, Jäger improved the method and compared this visualisation to some other methods such as Neighbour-Joining tree, multidimensional scaling and Principal  components analysis.  Some informative slides can be found here.

Below are some of the alternative visualisations (click for larger image).  Jäger points out that PCA is only really useful for the large language families, while CLANS is sensitive to local patterns.

The phoneme diagram is also updated in the talk slides, and actually resembles mine more (I assume Jäger means ‘Coronal/Dorsal’ instead of ‘dental’ below):

Also, Jäger took the phoneme distance measures produced by my variation of the algorithm above and computed the similarity matrix for Indo-European languages.  The Neighbor Joining tree computed from this has 51 false positives (17%), while the original algorithm produces 87 false positives.  So, far from being a coarse approximation, the distance measure has actually improved the performance!  The comparison is not entirely fair, however, since I ran my algorithm for twice as long as Jäger.  Still, this is an interesting development.

References

Jäger, G. (2012) Estimating and visualizing language similarities using weighted alignment and force-directed graph layout, to appear in Proceedings of LINGVIS & UNCLH, Workshop at EACL 2012, Avignon.

Wichmann, Søren, André Müller, Viveka Velupillai, Annkathrin Wett, Cecil H. Brown, Zarina Molochieva, Julia Bishoffberger, Eric W. Holman, Sebastian Sauppe, Pamela Brown, Dik Bakker, Johann-Mattis List, Dmitry Egorov, Oleg Belyaev, Matthias Urban, Harald Hammarström, Agustina Carrizo, Robert Mailhammer, Helen Geyer, David Beck, Evgenia Korovina, Pattie Epps, Pilar Valenzuela, and Anthony Grant. (2012). The ASJP Database (version 15).

Frickey T., Lupas A.N. (2004) CLANS: a Java application for visualizing protein families based on pairwise similarity. Bioinformatics 20:3702-3704

Bakker, D., Müller, A., Velupillai, V., Wichmann, S., Brown, C., Brown, P., Egorov, D., Mailhammer, R., Grant, A., & Holman, E. (2009). Adding typology to lexicostatistics: A combined approach to language classification Linguistic Typology, 13 (1), 169-181 DOI: 10.1515/LITY.2009.009

I recently was watching a debate between Richard Dawkins and Rowan Williams, the Archbishop of Canterbury, on the nature of the human species and its origin. To no one’s surprise, language was brought up when discussing human origins.  Specifically, recursive, productive language as a distinguishing marker of the human species. What may seem obvious to the evolutionary linguists here actually came with some interesting problems, from a biological perspective. As Dawkins discusses in the debate, evolution is rather difficult for the animal kingdom. Whereas for plants, there may be distinct moments at which one can point and say “Here is when a new species emerged!”, this identifiable moment is less overt for animals.  One key problem with determining the exact moment of a new species’ emergence is the question of interbreeding.

If we consider the development of a language (a system of communication with the aforementioned characteristics) to be a marker of the human species, then do we suppose at one point we have a child emerging with the ability to form a language with mute or animalistic parents? To whom would the child speak? If Dawkins is correct and language is partially rooted in a specific gene, we could theorize that the “first” human with the gene would thereby mate with proto-humans lacking the gene. All of this is, of course, very sketchy and difficult to elucidate, as even the theory that language is rooted in a gene can be disputed. The problem remains an integral one, not only for understanding the evolutionary origins, but as the philosophers in my pragmatics class would point out, it would also have significant bearing on ontological and ethical questions regarding human origins.

I do not hope to solve this entire issue in my senior thesis; however, I do hope to show the development of language less as a suddenly produced trait and more as a gradual process from a less developed system of communication to a more developed one. From a pragmatics point of view, the question might be, how do we jump the gap, so to speak, between the lesser developed systems of communication (conventionally, these include animal communication, natural meaning, etc.) and the fully fledged unique system of human language? Paul Grice, as one might discover in my handy dandy Wikipedia link above, proposed a distinction between natural meaning, which he defined as being a cause/effect indication and considered in terms of its factivity, and non-natural meaning, as a communicative action that must be considered in terms of the speaker’s intentions. Yet, as stated above, the question remains: how do we (evolutionarily) progress from natural meaning to non-natural meaning?

Not to overly simplify, but my answer rests in the question of what it means to mean something. I hope to show, in my subsequent posts, that an investigation into semantics, and, more specifically, a natural progression through a hierarchy of types of meaning, might shed light on this problem. In short, taking a look at the development of meaning, intent, and the qualifications for a language proper can shed light on how language developed into the complex, unique phenomenon we study today.  (Oh, and to satisfy the philosophers in my class, I may ramble occasionally about the implications for a philosophical conception of our species!)

Click here to play the QHImp Qhallenge!

It goes like this: there’s a tasty and terrified crustacean running around in a perspex box that has two possible ways of being opened by hungry octopuses. The experimenters capture a wild octopus (let’s call him Steve) from the harbour (which I’ll get back to in a minute), and they put it in a tank with the puzzle-box. After Steve stares hopelessly at the box for a while, it is then removed from the tank. Steve the kidnapped octopus then gets to watch a captive octopus in the next tank being presented with the same puzzle-box containing the delicious crab. Of course, the captive octopus has been confronted with the puzzle-box enough times that it has worked out a successful solution, and so opens the box like a pro. Steve is then presented with a crab in a puzzle-box again, except this time he goes straight for the crab using the same solution he just learned from the captive octopus. Here is a clip from the documentary, showing a trial of this experiment (NB: not the best quality).

It turns out that the papers on this go back as far as Fiorito & Scotto 1992, and it seems this was the first time observational learning had been demonstrated in invertebrates. The reason I’m interested in reporting this is because the documentary I watched explained another possible motivation/interpretation for Fiorito’s work that I can’t find in any of his actual papers. The octopuses used in the experiments were all caught from the harbour at Naples just before the experiments, which of course controlled for any prior experience with the puzzle boxes. But the results were reported as particularly interesting because the Naples harbour had been overfished and disrupted, resulting in an increase in marine predators that eat the small octopus vulgaris as well as fish that the octopuses themselves rely on. These harsher environmental conditions resulted in the octopuses being forced to inhabit a smaller space alongside each other. As a result, young octopuses were frequently exposed to, and even coexisted with, older octopuses. This is a weird situation for an octopus; they usually live solitary lives and never even meet their own mothers, who die of starvation while caring for the eggs (the fathers die within a few months of having mated). The only real interactions are mating, and conflicts between rival males while competing for a mate.

Toward the end of the documentary, the voiceover growled against some dramatic music about how the combination of observational learning capacities and increased predation pushing octopuses into groups meant that it was only a matter of time  before we’re overthrown by octopus vulgaris. This made me think of Dunbar’s “social brain hypothesis” for the emergence of language, and whether I really should prepare to welcome our new octopus overlords. Talking specifically about primates, Dunbar (1996) states that “[primates] in general exhibit two responses to increased predation: they grow physically bigger [or] they increase the size of their groups” (p.110). In order to maintain these groups, that are essential for survival in harsh ecological conditions, Dunbar suggests that standard primate grooming behaviour becomes too time consuming and costly in order to keep up with the rapidly expanding social group, creating a pressure for a more efficient method of bonding and communicating that allows the size of the group to continue increasing. Again with reference just to primate communication, Dunbar says “This [efficient mechanism] need not have involved any dramatic change, for as the studies by Seyfarth and Cheney have shown, primate vocalizations are already capable of conveying a great deal of social information and commentary.” (p.115) In addition to primates, and adding further credence to this idea, it’s been shown that the older Matriarchs of elephant groups make use of vocalisations to seemingly instruct their group on how to fend off lion attacks (McComb et al., 2011; previous Replicated Typo coverage here). It seems to me that the difference between the primates/elephants and the octopuses is that the former endeavour to actually enrich the environment from which their conspecifics extract information; they don’t just learn, they inform. As far as we can tell, there is no such communication like this – that is, enriching the environment in some way that helps other octopuses learn or survive – happening among the octopuses. That said, we know from mating displays and conflicts that cephalopods can communicate with chromatophore signalling. As an interesting aside that is reflective of their cognitive abilities and capacity for suffering, octopuses are treated as honorary vertebrates by UK animal testing laws.

It’s interesting enough that marine biologists (or at least those reporting on marine biologists) seem to have the same idea as Dunbar about the necessary preconditions for successful societies of animals, but why not let’s get wildly speculative? If (..!) existing in groups is in fact adaptive for these octopuses in the face of increased predation, and the competition between them for resources isn’t too great a counter-factor, it seems the only ingredient missing from an octopocalypse is the emergence of some cooperative behaviour. Someone should keep an eye on that harbour.

References

Dunbar, R. (1996) Grooming, Gossip, and the Evolution of Language. Harvard University Press: Cambridge, Massachusetts

Fiorito, G. & Scotto, P. (1992) “Observational learning in Octopus vulgaris” Science 256, 545-546.

McComb, K., Shannon, G., Durant, S., Sayialel, K., Slotow, R., Poole, J. & Moss, C. (2011) “Leadership in elephants: the adaptive value of age” Proceedings of the Royal Society B, published online.

Social dominance can be inferred in a couple of ways. Causal cues such as age, physical aggression and size can tell us about the dominance status of an individual quite intuitively, so we can make a sensible decision about whether or not we get into a scrap with them. Another way we can establish this is to look for direct realisations of dominance, such as who gets the banana if two hungry chimps both want it; chances are, little Pan Pipsqueak isn’t going to get a look in. In order to be useful, we also have to use this information to expect certain things from the individuals around us, so those representations have some property of stability across time that allows us to have those expectations. The question being explored in this paper is whether the representations we have are about the relationship between the two agents who want the banana, or the individual properties each of them has.

In a series of experiments using preferential looking time as a dependent measure, human infants (9 and 12 month olds) were exposed to videos of geometric figures exhibiting similar goal-directed behaviour. Then they would watch, say, a dominant triangle picking up the last figurative banana when the nondominant pentagon also wanted it. For expository purposes and posterity’s sake, I have constructed an artist’s impression of a dominant triangle and a subordinate pentagon in MSPaint (below, right):

A dominant triangle and subordinate pentagon (artist's impression).

I’m not just showing off my extraordinary artistic talent here; the good thing about these agents is that there are none of the cues like size or aggression that can give rise to the assignment of individual dominance properties. The task also doesn’t indicate anything similar; it’s just about who gets the desired object when there’s only one left. In other words, the goal-directed actions of two agents are in opposition. After seeing a triangle beat a pentagon to an object of ‘banana’ status, 12 month olds looked for longer when they were then presented with an incongruent trial where the pentagon gained over the triangle. 9 month olds (understandably?) couldn’t care less. So, on the basis of this social interaction alone, the 12 month olds were able to notice when something unexpected happened.

To rule out the possibility that this was just the result of some simple heuristic such as “when triangle and pentagon are present, triangle gets the object” and make sure the infants really were assigning some dominance, another experiment (with 12 and 15 month olds) showed the same test video of the two agents collecting little objects. This time, however, the preceding video was of the triangle dominating a little walled-in space that the pentagon also wanted to inhabit. The 12 month olds had no idea what was up, but the 15 month olds generalised from the first “get out of my room” interaction to the “I get the last banana” interaction. So, 15 month olds can extract, just from watching a social interaction, the dominance status of agents and can generalise that information to novel situations. So if a 15 month old watches you lose your favourite seat in front of the TV, they’ll also expect you to miss out on the last slice of pizza, because you’re a loser.

What we still don’t know is whether they think your belly is inherently yellow, or if you’re just a pushover when interacting with a particular person. Is it the relationship between the triangle and pentagon that the babies are tracking, or do they just give each agent some sort of dominance score? This was addressed in experiment 4, where infants were presented with two interactions: one between A and B, where A wins, and then another between B and C, where B wins. If the babies are assigning an individual value to each agent, they should have some sort of linear, transitive representation of dominance like A > B > C. If they’re then presented with a novel interaction between A and C, they would have the expectation that A will beat C. So if they stare in surprise at a trial where C wins, we know it’s violated that kind of expectation, and that they’re representing this stuff linearly – I.E. each agent has a dominance value. In contrast, if the infant is tracking the relations between agents, they can’t really have an expectation of what will happen when A and C both want a banana, because they’ve never seen C before. The results find that the infants look preferentially when they get an incongruent trial using agent pairs they have seen before – as we’d expect from the previous experiment. When they’re presented with a new “I get the last banana” interaction between A and C, however, there’s nothing startling about it when C wins – which means their expectations are not based on something like A > B > C.

The only tiny little harrumph I have about this result is that all it does is falsify the linear representation account. Though I think their account is absolutely right, it’d be nice to see something more predictive come out of the relation-representation hypothesis that is a little more falsifiable. But this result is pretty huge, and stands in contrast with what we know about social cognition in other animals like baboons (Cheney et al, 1995; Bergman et al, 2003), lemurs (Maclean et al., 2008) and even pigeons (Lazareva & Wasserman, 2012), who seem to employ this sort of hierarchical, transitive inference when presented with novel interactions. It may also muddy the waters a little when we want to make the appealing claim that, since language surely emerged in order to enable communication as we navigated a social environment, hierarchical social cognition gives rise to the processing of languagey things like hierarchical syntax or our semantic representation (Hamilton, 2005), which can be characterised as hierarchical (e.g. hyperonym > hyponym). If we consider the nature of the human social environment, though, it should seem more intuitive that something more reliable than simple transitive inference is necessary in order to successfully navigate through our interactions. Due to our prolific production of (and reliance on) culture, humans have a much more diverse range of social currencies, which correspond to values for things like money, intelligence, blackmail information, who your friends are, ad infinitum. That means it’s pretty reasonable that our social cognition needs new strategies in order to get by; we have a little more to consider than just who’s big and angry enough to get all the bananas.

References

Bergman, T., Beehner, J., Cheney, D. & Seyfarth, R. (2003) “Hierarchical Classification by Rank and Kinship in Baboons” Science 14(302), 1234-1236.

Cheney, D., Seyfarth, R. & Silk, J. (1995) “The response of female baboons (Papio cynocephalus ursinus) to anomalous social interactions: evidence for causal reasoning?” Journal of Comparative Psychology 109(2), 134-141.

Hamilton, D.L. (2005) Social Cognition: Key Readings (p. 104) Psychology Press

Lazareva, O. & Wasserman, E. (2012) “Transitive inference in pigeons: measuring the associative values of stimulus B and D” Behavioural Process 89(3), 244-255.

Maclean, E., Merritt, D. & Brannon, E.M. (2008) “Social complexity predicts transitive reasoning in prosimian primates” Animal Behaviour 76(2), 479-486.

Mascaro, O. & Csibra, G. (forthcoming) “Representation of stable dominance relations by human infants” Proceedings of the National Academy of Sciences

Each PhD project focuses on one of the approaches:

• iterated learning
• individual learning
• computational modeling

But it is envisaged that the PhD students interact in their research. The PhDs are paid a stipend of around 36.000 euros (before taxes) per year, and there is additional money for travel and research assistants.

More information, and instructions on how to apply can be found at http://ai.vub.ac.be/members/bart under “vacancies”.