Category Archives: Science

Screen Shot 2015-02-12 at 22.47.52

How spurious correlations arise from inheritance and borrowing (with pictures)

James and I have written about Galton’s problem in large datasets.  Because two modern languages can have a common ancestor, the traits that they exhibit aren’t independent observations.  This can lead to spurious correlations: patterns in the data that are statistical artefacts rather than indications of causal links between traits.

However, I’ve often felt like we haven’t articulated the general concept very well.  For an upcoming paper, we created some diagrams that try to present the problem in its simplest form.

Spurious correlations can be caused by cultural inheritance 

Gproblem2

Above is an illustration of how cultural inheritance can lead to spurious correlations.  At the top are three independent historical cultures, each of which has a bundle of various traits which are represented as coloured shapes.  Each trait is causally independent of the others.  On the right is a contingency table for the colours of triangles and squares.  There is no particular relationship between the colour of triangles and the colour of squares.  However, over time these cultures split into new cultures.  Along the bottom of the graph are the currently observable cultures.  We now see a pattern has emerged in the raw numbers (pink triangles occur with orange squares, and blue triangles occur with red squares).  The mechanism that brought about this pattern is simply that the traits are inherited together, with some combinations replicating more often than others: there is no causal mechanism whereby pink triangles are more likely to cause orange squares.

Spurious correlations can be caused by borrowing

Gproblem_HorizontalB

Above is an illustration of how borrowing (or areal effects or horizontal cultural inheritance) can lead to spurious correlations.  Three cultures (left to right) evolve over time (top to bottom).  Each culture has a bundle of various traits which are represented as coloured shapes.  Each trait is causally independent of the others.  On the right is a count of the number of cultures with both blue triangles and red squares.  In the top generation, only one out of three cultures have both.  Over some period of time, the blue triangle is borrowed from the culture on the left to the culture in the middle, and then from the culture in the middle to the culture on the right.  By the end, all languages have blue triangles and red squares.  The mechanism that brought about this pattern is simply that one trait spread through the population: there is no causal mechanism whereby blue triangles are more likely to cause red squares.

A similar effect would be caused by a bundle of causally unrelated features being borrowed, as shown below.

Gproblem_Horizontal

Screen Shot 2015-01-27 at 09.59.37

Tone and Humidity: FAQ

Everett, Blasi & Roberts (2015) review literature on how inhaling dry air affects phonation, suggesting that lexical tone is harder to produce and perceive in dry environments.  This leads to a prediction that languages should adapt to this pressure, so that lexical tone should not be found in dry climates, and the paper presents statistical evidence in favour of this prediction.

Below are some frequently asked questions about the study (see also the previous blog post explaining the statistics).

Continue reading

call for papers

Reminder for upcoming conferences

The deadline is approaching for several relevant call for papers:

At this year’s International Congress of Phonetic Sciences in Glasgow there will be a special interest group on the Evolution of our phonetic capabilities. It will focus on the interaction between biological and cultural evolution and encourages work from different modalities too. The deadline is 16th Feb. The call for papers is here.

There’s also a special discussant session on Sound change and speech evolution at ICPhS headed by Andy Wedel. The deadline for the actual conference is 1st Feb. Call for Papers here.

The next event in the ways to (proto)language conference is being held in Rome! The deadline is also 1st Feb. Call for Papers here.

This year’s CogSci is being organised by the guys at Cognitive and Information Sciences at the University of California in Merced, who do some great stuff related to language evolution. The deadline is 1st Feb as well, and the call for paper is here. 

Happy submitting!

saussure-with-money

The Vocal Iconicity Challenge!

Do you fancy the prospect of putting your communication skills to the test and winning $1000? If so, you should probably go and check out The Vocal Iconicity Challenge: http://sapir.psych.wisc.edu/vocal-iconicity-challenge/

Devised by Gary Lupyan and Marcus Perlman, of the University of Wisconsin-Madison, the aim of the game is to devise a system of vocalizations to communicate a set of Paleolithic-relevant meanings. The team whose vocalizations are guessed most accurately will be crowned the Vocal Iconicity Champion (and win the $1000 Saussure Prize!). More information is on their website.

Screen Shot 2014-12-18 at 21.16.38

Evolve an App Name

Edit: The results are out!

I’m working with the Language in Interaction project to create an App game about linguistic diversity.  It’s a game where you listen to several recordings of people talking and have to match the ones who are speaking the same language.  It’s quite a lot like the Great Language Game, but we’re using many lesser-known languages from the DOBES archive.

But first – we need a name.  Help us create one with the power of Iterated Learning!

Click to take part in our 1-minute experiment to evolve an app name.

We’ll throw some app names at you, you try to remember them, then we throw your names at someone else.

Here’s a screenshot of the App in development:

Drag2_demo

(P.S.: I’ve done this kind of thing before to evolve a band name)

A heatmap showing an example language from one of the Shape Different condition chains at Generation 3. Vertically we have the labels produced and horizontally we have the stimuli. Colour represent token frequency that has been rescaled from 0-1 (darker representing higher frequency).

Languages adapt to their contextual niche (Winters, Kirby & Smith, 2014)

ResearchBlogging.orgLast week saw the publication of my latest paper, with co-authors Simon Kirby and Kenny Smith, looking at how languages adapt to their contextual niche (link to the OA version and here’s the original). Here’s the abstract:

It is well established that context plays a fundamental role in how we learn and use language. Here we explore how context links short-term language use with the long-term emergence of different types of language systems. Using an iterated learning model of cultural transmission, the current study experimentally investigates the role of the communicative situation in which an utterance is produced (situational context) and how it influences the emergence of three types of linguistic systems: underspecified languages (where only some dimensions of meaning are encoded linguistically), holistic systems (lacking systematic structure) and systematic languages (consisting of compound signals encoding both category-level and individuating dimensions of meaning). To do this, we set up a discrimination task in a communication game and manipulated whether the feature dimension shape was relevant or not in discriminating between two referents. The experimental languages gradually evolved to encode information relevant to the task of achieving communicative success, given the situational context in which they are learned and used, resulting in the emergence of different linguistic systems. These results suggest language systems adapt to their contextual niche over iterated learning.

Background

Context clearly plays an important role in how we learn and use language. Without this contextual scaffolding, and our inferential capacities, the use of language in everyday interactions would appear highly ambiguous. And even though ambiguous language can and does cause problems (as hilariously highlighted by the ‘What’s a chicken?’ case), it is also considered to be communicatively functional (see Piantadosi et al., 2012).  In short: context helps in reducing uncertainty about the intended meaning.

If context is used as a resource in reducing uncertainty, then it might also alter our conception of how an optimal communication system should be structured (e.g., Zipf, 1949). With this in mind, we wanted to investigate the following questions: (i) To what extent does the context influence the encoding of features in the linguistic system? (ii) How does the effect of context work its way into the structure of language?  To get at these questions we narrowed our focus to look at the situational context: the immediate communicative environment in which an utterance is situated and how it influences the distinctions a speaker needs to convey.

Of particular relevance here is Silvey, Kirby & Smith (2014): they show that the incorporation of a situational context can change the extent to which an evolving language encodes certain features of referents. Using a pseudo-communicative task, where participants needed to discriminate between a target and a distractor meaning, the authors were able to manipulate which meaning dimensions (shape, colour, and motion) were relevant and irrelevant in conveying the intended meaning. Over successive generations of participants, the languages converged on underspecified systems that encoded the feature dimension which was relevant for discriminating between meanings.

The current work extends upon these findings in two ways: (a) we added a communication element to the setup, and (b) we further explored the types of situational context we could manipulate.  Our general hypothesis, then, is that these artificial languages should adapt to the situational context in predictable ways based on whether or not a distinction is relevant in communication.

Continue reading

IMGP8152rd2

John Lawler on Generative Grammar

From a Facebook conversation with Dan Everett (about slide rules, aka slipsticks, no less) and others:

The constant revision and consequent redefining and renaming of concepts – some imaginary and some very obvious – has led to a multi-dimensional spectrum of heresy in generative grammar, so complex that one practically needs chromatography to distinguish variants. Babel comes to mind, and also Windows™ versions. Most of the literature is incomprehensible in consequence – or simply repetitive, except it’s too much trouble to tell which.

–John Lawler
IMGP9487rdXFdEq

Why I Abandoned Chomskian Linguistics, with Links to 2 FB Discussions with Dan Everett

It wasn’t a matter of deep and well-thought princple. It was simpler than that. Chomsky’s approach to linguistics didn’t have the tools I was looking for. Let me explain.

* * * * *

Dan Everett’s kicked off two discussions on Facebook about Chomksy. This one takes Christina Behme’s recent review article, A ‘Galilean’ science of language, as its starting point. And this one’s about nativism, sparked by Vyv Evans’ The Language Myth.

* * * * *

I learned about Chomsky during my second year as an undergraduate at Johns Hopkins. I took a course in psycholinguistics taught by James Deese, known for his empirical work on word associations. We read and wrote summaries of classic articles, including Lee’s review of Syntactic Structures and Chomsky’s review of Skinner’s Verbal Behavior. My summary of one of them, I forget which, prompted Deese to remark that my summary was an “unnecessarily original” recasting of my argument.

That’s how I worked. I tried to get inside the author’s argument and then to restate it in my own words.

In any event I was hooked. But Hopkins didn’t have any courses in linguistics let alone a linguistics department. So I had to pursue Chomsky and related thinkers on my own. Which I did over the next few years. I read Aspects, Syntatic Structures, Sound Patterns of English (well, more like I read at that one), Lenneberg’s superb book on biological foundations (with an appendix by Chomsky), found my way to generative semantics, and other stuff. By the time I headed off to graduate school in English at the State University of New York at Buffalo I was mostly interested in that other stuff.

I became interested in Chomsky because I was interested in language. While I was interested in language as such, I was a bit more interested in literature and much of my interest in linguistics followed from that. Literature is made of language, hence some knowledge of linguistics should be useful. Trouble is, it was semantics that I needed. Chomsky had no semantics and generative semantics looked more like syntax.

So that other stuff looked more promising. Somehow I’d found my way to Syd Lamb’s stratificational linguistics. I liked that for the diagrams, as I think diagrammatically, and for the elegance. Lamb used the same vocabulary of structural elements to deal with phonology, morphology, and syntax. That made sense to me. And the s work within his system actually looked like semantics, rather than souped up syntax, though there wasn’t enough of it. Continue reading

Language as a multimodal phenomenon

The issue of multimodality has become a widely discussed topic in several branches of linguistics and especially in research on the evolution of language. Now, a special issue of the “Philosophical Transactions of the Royal Society B” has been dedicated to “Language as a multimodal phenomenon”. The issue, edited by Gabriella Vigliocco, Pamela Perniss, and David Vinson, features a variety of interesting papers by outstanding scholars from different fields such as gesture research, signed language research, neurolinguistics, and evolutionary linguistics.

For example, Susan Goldin-Meadow discusses “what the manual modality reveals about language, learning and cognition”, arguing that, in child language acquisition, manual gestures “precede, and predict, the acquisition of structures in speech”.

Ulf Liszkowski addresses the question of how infants communicate before they have acquired a language, and Aslı Özyürek reviews neuroscientific findings on “Hearning and seeing meaning in speech and gesture”. Jeremy Skipper discusses “how auditory cortex hears context during speech perception”, and Stephen Levinson and Judith Holler, in a paper entitled “The origin of human multi-modal communication”,  talk about “the different roles that the different modalities play in human communication, as well as how they function as one integrated system despite their different roles and origins.”

Martin Sereno, in his opinion piece on the “Origin of  symbol-using systems”, argues that we have to distinguish “the origin of a system capable of evolution from the subsequent evolution that system becomes capable of”. According to Sereno,

“Human language arose on a substrate of a system already capable of Darwinian evolution; the genetically supported uniquely human ability to learn a language reflects a key contact point between Darwinian evolution and language. Though implemented in brains generated by DNA symbols coding for protein meaning, the second higher-level symbol-using system of language now operates in a world mostly decoupled from Darwinian evolutionary constraints.”

Padraic Monaghan, Richard C. Shillcock, Morten H. Christiansen, and Simon Kirby address the question “How arbitrary is language?” Drawing on a large-scale corpus analysis, they show that

“sound–meaning mappings are more systematic than would be expected by chance. Furthermore, this systematicity is more pronounced for words involved in the early stages of language acquisition and reduces in later vocabulary development.”

Mutsumi Imai and Sotaro Kita propose a “sound symbolism bootstrapping hypothesis for language acquisition and language evolution”, arguing that “sound symbolism helps infants and toddlers associate speech sounds with their referents to establish a lexical representation” and that sound symbolism might be deeply related to language evolution.

Karen Emmorey discusses the role of iconicity in sign language grammar and processing, and in the final paper, Pamela Perniss and Gabriella Vigliocco argue that ” iconicity in face-to-face communication (spoken and signed) is a powerful vehicle for bridging between language and human sensori-motor experience, and, as such, iconicity provides a key to understanding language evolution, development and processing.”

The special issue is available here. Some of the papers are open access, all others can be accessed freely until October 19th ( User name: language; Password: tb1651 – since this information was distributed by the Royal Sociaty via several mailing lists, I guess I’m free to share it here).

 

wales

Vyv Evans: The Human Meaning-Making Engine

If you read my last post here at Replicated Typo to the very end, you may remember that I promised to recommend a book and to return to one of the topics of this previous post. I won’t do this today, but I promise I will catch up on it in due time.

What I just did – promising something – is a nice example for one of the two functions of language which Vyvyan Evans from Bangor University distinguished in his talk on “The Human Meaning-Making Engine” yesterday at the UK Cognitive Linguistics Conference. More specifically, the act of promising is an example for the interactive function of language, which is of course closely intertwined with its symbolic function. Evans proposed two different sources for this two functions. The interactive function, he argued, arises from the human instinct for cooperation, whereas meaning arises from the interaction between the linguistic and the conceptual system. While language provides the “How” of meaning-making, the conceptual system provides the “What”. Evans used some vivid examples (e.g. this cartoon exemplifying nonverbal communication) to make clear that communication is not contingent on language. However, “language massively amplifies our communicative potential.” The linguistic system, he argued, has evolved as an executive control system for the conceptual system. While the latter is broadly comparable with that of other animals, especially great apes, the linguistic system is uniquely human. What makes it unique, however, is not the ability to refer to things in the world, which can arguably be found in other animals, as well. What is uniquely human, he argued, is the ability to symbolically refer in a sign-to-sign (word-to-word) direction rather than “just” in a sign-to-world (word-to-world) direction.  Evans illustrated this “word-to-word” direction with Hans-Jörg Schmid’s (e.g.  2000; see also here)  work on “shell nouns”, i.e. nouns “used in texts to refer to other passages of the text and to reify them and characterize them in certain ways.” For instance, the stuff I was talking about in the last paragraph would be an example of a shell noun.

According to Evans, the “word-to-word” direction is crucial for the emergence of e.g. lexical categories and syntax, i.e. the “closed-class” system of language. Grammaticalization studies indicate that the “open-class” system of human languages is evolutionarily older than the “closed-class” system, which is comprised of grammatical constructions (in the broadest sense). However, Evans also emphasized that there is a lot of meaning even in closed-class constructions, as e.g. Adele Goldberg’s work on argument structure constructions shows: We can make sense of a sentence like “Someone somethinged something to someone” although the open-class items are left unspecified.

Constructions, he argued, index or cue simulations, i.e. re-activations of body-based states stored in cortical and subcortical brain regions. He discussed this with the example of the cognitive model for Wales: We know that Wales is a geographical entity. Furthermore, we know that “there are lots of sheep, that the Welsh play Rugby, and that they dress in a funny way.” (Sorry, James. Sorry, Sean.) Oh, and “when you’re in Wales, you shouldn’t say, It’s really nice to be in England, because you will be lynched.”

On a more serious note, the cognitive models connected to closed-class constructions, e.g. simple past -ed or progressive -ing, are of course much more abstract but can also be assumed to arise from embodied simulations (cf. e.g. Bergen 2012). But in addition to the cognitive dimension, language of course also has a social and interactive dimension drawing on the apparently instinctive drive towards cooperative behaviour. Culture (or what Tomasello calls “collective intentionality”)  is contigent on this deep instinct which Levinson (2006) calls the “human interaction engine”. Evans’ “meaning-making engine” is the logical continuation of this idea.

Just like Evans’ theory of meaning (LCCM theory), his idea of the “meaning-making engine” is basically an attempt at integrating a broad variety of approaches into a coherent model. This might seem a bit eclectic at first, but it’s definitely not the worst thing to do, given that there is significant conceptual overlap between different theories which, however, tends to be blurred by terminological incongruities. Apart from Deacon’s (1997) “Symbolic Species” and Tomasello’s work on shared and joint intentionality, which he explicitly discussed, he draws on various ideas that play a key role in Cognitive Linguistics. For example, the distinction between open- and closed-class systems features prominently in Talmy’s (2000) Cognitive Semantics, as does the notion of the human conceptual system. The idea of meaning as conceptualization and embodied simulation of course goes back to the groundbreaking work of, among others, Lakoff (1987) and Langacker (1987, 1991), although empirical support for this hypothesis has been gathered only recently in the framework of experimental semantics (cf. Matlock & Winter forthc. – if you have an account at academia.edu, you can read this paper here). All in all, then, Evans’ approach might prove an important further step towards integrating Cognitive Linguistics and language evolution research, as has been proposed by Michael and James in a variety of talks and papers (see e.g. here).

Needless to say, it’s impossible to judge from a necessarily fairly sketchy conference presentation if this model qualifies as an appropriate and comprehensive account of the emergence of meaning. But it definitely looks promising and I’m looking forward to Evans’ book-length treatment of the topics he touched upon in his talk. For now, we have to content ourselves with his abstract from the conference booklet:

In his landmark work, The Symbolic Species (1997), cognitive neurobiologist Terrence Deacon argues that human intelligence was achieved by our forebears crossing what he terms the “symbolic threshold”. Language, he argues, goes beyond the communicative systems of other species by moving from indexical reference – relations between vocalisations and objects/events in the world — to symbolic reference — the ability to develop relationships between words — paving the way for syntax. But something is still missing from this picture. In this talk, I argue that symbolic reference (in Deacon’s terms), was made possible by parametric knowledge: lexical units have a type of meaning, quite schematic in nature, that is independent of the objects/entities in the world that words refer to. I sketch this notion of parametric knowledge, with detailed examples. I also consider the interactional intelligence that must have arisen in ancestral humans, paving the way for parametric knowledge to arise. And, I also consider changes to the primate brain-plan that must have co-evolved with this new type of knowledge, enabling modern Homo sapiens to become so smart.

 

References

Bergen, Benjamin K. (2012): Louder than Words. The New Science of How the Mind Makes Meaning. New York: Basic Books.

Deacon, Terrence W. (1997): The Symbolic Species. The Co-Evolution of Language and the Brain. New York, London: Norton.

Lakoff, George (1987): Women, Fire, and Dangerous Things. What Categories Reveal about the Mind. Chicago: The University of Chicago Press.

Langacker, Ronald W. (1987): Foundations of Cognitive Grammar. Vol. 1. Theoretical Prerequisites. Stanford: Stanford University Press.

Langacker, Ronald W. (1991): Foundations of Cognitive Grammar. Vol. 2. Descriptive Application. Stanford: Stanford University Press.

Levinson, Stephen C. (2006): On the Human “Interaction Engine”. In: Enfield, Nick J.; Levinson, Stephen C. (eds.): Roots of Human Sociality. Culture, Cognition and Interaction. Oxford: Berg, 39–69.

Matlock, Teenie & Winter, Bodo (forthc): Experimental Semantics. In: Heine, Bernd; Narrog, Heiko (eds.): The Oxford Handbook of Linguistic Analysis. 2nd ed. Oxford: Oxford University Press.

Schmid, Hans-Jörg (2000): English Abstract Nouns as Conceptual Shells. From Corpus to Cognition. Berlin, New York: De Gruyter (Topics in English Linguistics, 34).

Talmy, Leonard (2000): Toward a Cognitive Semantics. 2 vol. Cambridge, Mass: MIT Press.