8a. Pinker, S. & Bloom, P. (1990). Natural language and natural selection
What's wrong and right about Steve Pinker's views on language evolution? And what was so special about language that the capacity to acquire it became evolutionarily encoded in the brains of our ancestors – and of no other surviving species – about 300,000 years ago? (It gave our species a unique new way to acquire categories, through symbolic instruction rather than just direct sensorimotor induction.)
Reading: Pinker, S. & Bloom, P. (1990). Natural language and natural selection. Behavioral and Brain Sciences 13(4): 707-784.
Instructions for commenting: Quote the passage on which you are commenting (use italics, indent). Comments can also be on the comments of others. Make sure you first edit your comment in another text processor, because if you do it directly in the blogger window you may lose it and have to write it all over again.
“Language is a complex biological adaptation to communicate information, shaped by natural selection for the transmission of knowledge.” (Pinker & Bloom, 1990)
ReplyDeleteWhat stuck out to me the most about the reading is the argument that language is a biological adaptation instead of a cultural invention. Language being a biological adaptation would suggest it evolved through natural selection because it helped increase the survival of our ancestors. They theorize like how the eye evolved as a natural selection mechanism to process light; language emerged to process and share information. I can see where they're coming from, especially in terms of the value that language holds in coordinating tasks and passing along knowledge.
The problem is just how unique this capacity was. No other species appears to have developed the same form of open-ended communication. Pinker and Bloom discuss the adaptive significance of language structure; however, I believe what is even more extraordinary is the way language permits us to learn categories and concepts in an indirect way, not just through experience but through words and descriptions. That jump from sensorimotor learning to symbolic learning feels like a significant defining moment in the evolutionary history that created what we now know as human cognition.
I agree that the case for language as a biological adaptation is highly convincing. Its ability to enable social coordination, teaching, persuasion, and even mating displays demonstrates how language could have increased the fitness of our ancestors. Pinker and Bloom emphasize that the complexity and universality of human language make it a species-specific adaptation, not a culturally learned skill. They attribute this capacity to specialized biological features such as the human brain, vocal tract, and perceptual systems for decoding and segmenting speech. While the paper argues that language evolved to transmit propositional structures between individuals, the authors clarify that it is a system shaped by cognition rather than one that transforms it. As they note, “Human knowledge and reasoning is couched in a ‘language of thought’ that is distinct from external languages such as English or Japanese… whose symbols pertain to people, objects, and events, the categories they belong to, their distribution in space and time, and their causal relations to one another” (p. 712). In this way, they differentiate between mental language, which organizes thought internally, and spoken language, which evolved to communicate those thoughts to others.
DeleteWhat I found particularly striking in Pinker and Bloom’s paper is how they defend language not just as a tool for communication, but as a biological design shaped for transmitting abstract structures with efficiency and precision. This idea deepens your point, Lorena, about language being an adaptation. This structure allows humans to communicate about hypothetical states, something no other species can do. It’s not only about coordination or survival, but about how our minds evolved to share internal representations of the world. That link between biological adaptation and the architecture of thought makes language feel less like a cultural invention and more like the evolutionary foundation of human cognition.
DeleteWhat’s interesting is if we take Pinker and Bloom’s definition, AI systems like ChatGPT don’t have language in the human sense, they model it. Their language isn’t grounded in cognition or social interaction, it’s a reconstruction of how humans use words. So while “language model” is a technically accurate term in computer science (referring to systems that predict sequences of words), it doesn’t qualify as human language in the evolutionary or cognitive sense that Pinker and Bloom describe.
I too was trying to compare Pinker and Bloom's definition of language to that of how AI uses language. As we know, artificial intelligence systems can convincingly mimic human linguistic patterns, but lack genuine understanding. Due to the lack of understanding and real connection of words and their referent, language models fail to meet the definition of language from the evolutionary or cognitive view.
Delete***EVERYBODY PLEASE NOTE: I REDUCED THE MINIMUM NUMBER OF SKYWRITINGS. BUT THE READINGS ARE **ALL** RELEVANT TO AN OVERALL UNDERSTANDING OF THE COURSE. SO, EVEN IF YOU DO NOT DO A SKYWRITING ON ALL OF THEM, AT LEAST FEED EACH READING YOU DO NOT READ TO CHATGPT AND ASK IT FOR A SUMMARY, SO YOU KNOW WHAT THE READING SAID — OTHERWISE YOU WILL NOT HAVE A COMPLETE GRASP OF THE COURSE TO INTEGRATE AND INTERCONNECT FOR THE FINAL EXAM.***
DeleteLorena, yes, the capacity for natural language evolved, through Darwinian and Baldwinian evolution, and it in turn provided a new way of communicating as well as learning (not just categories but even motor skills) that is unique to our species (instead just pointing, showing or imitating). The biggest evolutionary challenge is how propositionality evolved. What is that? And do you have any ideas about how and why propositionality began?
Emily P&B's "language of thought" as distinct from natural language itself is a little speculative (and originates from Fodor, and perhaps partly Chomsky, but not necessarily in the same sense: ask GPT). And natural language is not necessarily vocal (and "Stevan Says" that it did not start in the vocal modality: why not?) but once propositionality began it migrated: why?
Ayla, P&B get ahead of themselves with "communicating hypothetical states". How do you get from a non-propositional primate to "communicating hypothetical states"? Doesn't that already require language? Ditto for sharing 'internal "representations" of the world (WW). Too quick as an explanation of how language began rather than the later things it turned out to be useful for.
LLMs are a challenging case. They can communicate with humans who have and can communicate in natural language. But do the LLMs understand language, hence "have" it? Or are they doing something else?
Kaelyn "AI" is becoming a WW since it refers willy nilly to ungrounded both (1) LLMs with huge databases of words produced by grounded human heads and to (2) a grounded sensorimotor robot (T3) that can interact with us humans indistinguishably (like our class robot, Anne-Sophie). Can you point out some the many issues in this course that that conflates (and how?).
Instructor I will try to tackle two of your three questions on propositionality from the beginning of your comment. (1) skipped. (2) what is propositionality? , it is the ability of language to express propositions (which are : subject + predicate + truth value). (3) ideas about how/why propositionality began? , this links to direct and indirect grounding. In a proposition, the subject is typically a content word and the predicate gives the distinguishable features of the subject. In that sense, propositionality began to share valuable knowledge (i.e. reduce uncertainty) from someone who learned categories through trial and error (sensorimotor direct learning) to someone who hasn't yet (but will now be learning through indirect verbal learning).
DeletePinker and Bloom (1990) write, “Language is an adaptation to the communication of complex propositions, a system of great intricacy and functional design” (p. 708). This means language evolved because it allowed humans to share more complex thoughts and ideas, something no other species can do. Pinker and Bloom argue that language is a biological adaptation that we human acquired brought abut by natural selection and not simply as a by-product of intelligence or a higher level cognitive capacity. What’s really interesting is how they link language itself to symbolic or more complex thought: humans could now pass on categories and content through words, not just through sensory experience. This ability to 'learn by instruction/words' rather than trial and error may have given our ancestors a massive survival advantage! This could explain why the capacity for language may have become hardwired in human brains but not in any other species, I wonder?
ReplyDeleteHey Rachel,
DeleteGreat points!! I do think you're right that language and complex thought likely co-evolved and think you articulate the paper's argument well. I think it's really interesting to also think about how other forms of animal communication, like warning calls for example, are specific to the present context or time. For us humans, we are able to discuss and communicate about ideas and concepts not limited to the immediate. Perhaps it is this ability to think symbolically that allowed for this evolution? I think your question regarding hardwiring in human brains is quite fascinating.
I’m just now understanding what was meant by the term “Just-So” stories, which describes adaptive functions that are invented in hindsight, such as our noses were made to carry spectacles . This is differentiated from complex functional design (such as eyes or language) which can only be explained through natural selection, as these intricate things cannot be explained away as a byproduct. I think this paper does a great job of outlining this difference.
ReplyDeleteI agree with your point on how Pinker and Bloom did a good job of emphasizing the difference. I like how they make sure their argument doesn’t become a “Just-So” story by explaining that language can’t just be a lucky by-product of intelligence, but something that had to evolve specifically because of the clear advantage it gives us in communicating complex propositions. What makes their argument strong is that they treat language like any other biological system (they show it is shaped by purpose similarly to how eyes are shaped for vision). They argue that complexity and universality of grammar are too functionally complex to be explained by mere coincidence. They’re tapping into discussion of why language exists. Natural selection favored minds that could turn grounded concepts into shared symbols for survival.
DeletePinker and Bloom argue that human language evolved through natural selection rather than as a random byproduct of other abilities. They challenge ideas from Chomsky and Gould by showing that grammar's complexity, like how we map thoughts onto spoken words, has all the signs of something that evolved because it helped our ancestors survive and communicate better.
ReplyDeleteOne thing that stood out to me was their point about arbitrary grammatical rules being useful because everyone in a community needs to use the same system to understand each other. This reminds me of the symbol grounding problem, which asks how abstract words get their meanings. I feel like this can be connected as the paper speculates that language evolved to express thoughts we already had and therefore, as per the solution explored by Harnad, our brains must have already figured out how to connect concepts to the real world before language even existed through categorization.
“Language was not invented by some groups and spread to others like agriculture or the alphabet. All languages are complex computational systems using the same basic kinds of rules and representations.”
ReplyDeleteWhat I found striking about Pinker and Bloom’s claim is how they frame language as a universal computational system rather than a cultural tool. The evolved adaptation is not language as a specific set of words, but the mental capacity to convert thought (internal representations) into symbolic, rule-governed expressions that others can understand. The universality of language might also explain why no other species exhibits the combinatorial and hierarchical structure of human language, even though they may have rich signalling systems (like bird songs or primate calls). Language, then, may have been the evolutionary point where thought itself became transmissible and, perhaps, where cognition gained an interface.
Hey Maya! I really like your point that language wasn’t invented by one group and then copied by others. It’s very interesting to me that every community came up with its own language, yet they all follow the same deep structure of rules and symbols. That really puts it into perspective for me about how the ability for language is something built into our brains. I also wonder how different environments shaped the way languages sound. Maybe all that variation just shows evolution playing around with the same blueprint.
DeleteI was particularly interested by the discussion of how languages evolved in a manner which was biologically convergent but culturally differentiated, which preserves the notion of a universal basis for language whilst explaining the differences in linguistic construction.
ReplyDeleteThey first suggest that it is common for structures to evolve to serve (potentially or actually) a small number of definite functions in different environments – in this sense, the same biological disposition for a general form of behaviour x may lead to a different expression depending on the needs of the environment or the dominant customs within it. This is reflected in linguistics by the very small number of possible pairings of grammatical devices, as well as the fact that linguistic diversity often reflects very small mental grammar differences.
It is further possible to suggest that what is mainly common is the ability to learn culturally relevant labels, but that it is from then almost impossible to develop a universal innate code, for very small differences in phenomenology may occur.
Something that interested me about Pinker and Bloom’s writing is the distinction between evolution and acquisition. Evolution selected for a functional communicative system, but by the time a new child is born, arbitrary specifics of the language have already been decided. The child then has to learn that particular kind of language that was predetermined for them. So, evolution coded for standards in language driven by their advantages, leaving the individual powerless to decide on any functional parts. Pinker and Bloom think that evolution and language not only can differ, but they must differ.
ReplyDeleteI agree that this is one of the more fascinating parts of this paper; once this distinction between ontogeny and phylogeny is made, it seems so obvious! Why should the way in which language evolved mirror the way an individual learns language? These mechanisms are so separate, it's led me to wonder which of our learned traits, if any, developed similarly on an individual level as they did in our species.
DeleteAs I explore this, I'm finding that this line is quite blurred. How could we know if a given behavior arose as a result of its immediate functional benefit, or as a reflection of an evolutionarily engrained tendency? This is especially difficult because we know from Cauchoix and Chaine's paper that evolution gives rise to mechanisms, not behaviors. This means that what looks like a functional adaptation to a present day problem might just be an extraordinary implementation of our evolutionary mechanisms within us.
I'm sure there are ways to make this line less blurry, but I think the problem is interesting. I apologize for this stream of consciousness-like response
I agree with most on the fact that the argument for the evolution of language being a product of Darwinian evolution seems convincing. Nevertheless, I think there may be more to it than Pink and Bloom suggest. In this paper, they argue that language must have evolved because it increased our species’ chances of survival and must have, therefore, been “picked” by natural selection. They even compare it to the eye on grounds of necessity and complexity. However, the process of natural selection is extremely slow, spanning several thousands to several millions of years. The capacity for language the way it is employed by humans evolved far quicker. Our ability to detect, abstract, and learn categories through sensorimotor interaction evolved hundreds of millions of years ago. Yet, it is theorized that our capacity to learn categories through verbal descriptions evolved only a few hundreds of thousands of years ago. Now, this doesn’t deconstruct Pinker and Bloom’s argument entirely, but it does raise the question of whether it is plausible that language evolved through natural selection alone. Could there have been an interaction with some other process (similarly to how cognition was proven to not be a result of computation alone)?
ReplyDeleteI really like how you brought up the timescale issue, I also found Pinker and Bloom’s comparison between language and the eye persuasive at first, but you’re right that the speed of linguistic evolution doesn’t totally line up with how slow natural selection usually works. If they are framing language as a completely biological adaptation, I feel like they might be simplifying it. Maybe what happened is the brain evolved certain cognitive abilities and then culture used those abilities to rapidly expand them into extensive language systems, explaining why language can change so fast among human groups (so the evolution aspect was the ability to change and develop language). So I agree that natural selection probably played a huge role in getting the foundations in place, but the global diversity and speed of language development might point to culture playing a larger role than argued by Pinker and Bloom.
DeleteI think the paper’s job of likening language’s complex propositional design with the organs in a human body easily puts into perspective the argument that language is a product of natural selection. The shift from a cultural or even evolutionary byproduct to a more biologically focused lens radically changes the way that I think of language. I do find it interesting though that they do make the distinction between an evolutionary byproduct—which they compare to a spandrel—and as the actual evolution itself. Which the distinction actually seems to tie into the ‘just-so’ stories, where language does not exist to fit an already pre-existing perfectly ‘language-shaped’ hole create by other evolved cognitive capacities, but rather it evolved to be something in its own right.
ReplyDeleteI thought your point about the organ analogy was intersting and accuratee. It really helps make sense of how Pinker and Bloom see language as something shaped by evolution, not just a cultural invention. I also liked that you brought up the “spandrel” idea. The authors basically say that even if some parts of grammar seem arbitrary, they still work because everyone shares the same system, similar to how random design quirks can still be useful once they’re standardized. I agree that it changes how we think about language. I think it changes not as a side effect of intelligence, but as its own biological adaptation for communication. It’s very interesting to think of grammar like an evolved tool our brains built for sharing ideas.
Delete“Human language meets these criteria: Grammar is a complex mechanism tailored to the transmission of propositional structures through a serial interface.”
ReplyDeleteI really liked this, it captures how they frame grammar not as an abstract set of arbitrary rules, but as something functional.. an evolved biological design that serves a clear purpose in communication. It ties language to survival and cognition, showing that grammar itself might have been shaped by the same selective pressures that shaped other adaptive systems like vision. If grammar is evolved through natural selection for communication, what kinds of selective pressures would have shaped it the most? Social cooperation, information sharing, mating signals, or something else?
This comment has been removed by the author.
ReplyDelete“[Chomsky and Gould] have repeatedly suggested that language may not be the product of natural selection, but a side effect of other evolutionary forces such as an increase in overall brain size and constraints of as yet unknown laws of structure and growth.”
ReplyDeleteAccording to Gould, language is a by-product of the neural growth and complexity of the brain and is a spandrel. Gould thinks language originally emerged as an exaptation of reasoning and internal thought, which later was specialized for communication. This point made me question his argument, as internal thought would certainly be able to be performed without language. I personally do not think it could be developed “only” as an exaptation of another feature of our brain, due to the specific mechanisms that arose in most (if not all) languages across the word such as phonetics and syntax. If language really was a by-product of another neural development, then those features emerging in most languages worldwide would be an incredibly unlikely coincidence. We use language to communicate with others, especially to communicate our thoughts, feelings and knowledge. Language is invaluable in those moments, because you communicate to someone who does not know your thoughts or what you are prepared to communicate to them. You transmit them information and use language to frame that information as an understandable sentence, by using context. That information is valuable only to the person you are communicating with, since you know the context while formulating the thought. Following language development, the rules we use to establish a context in a conversation is mostly the same across all languages, which is evidence of the evolutionary aspect of language in my opinion.
“Language is a complex biological adaptation to communicate information, shaped by natural selection for the transmission of knowledge.” (Pinker & Bloom, 1990)
ReplyDeleteI agree that Pinker and Bloom make a convincing case that language evolved through Darwinian natural selection, but their argument also seems compatible with Baldwinian evolution. If early humans who could learn communication systems more quickly or effectively had survival advantages, then genes supporting faster or more flexible learning would be naturally selected for. Over time, this could lead to a biologically evolved capacity for language learning itself, a Darwinian outcome emerging through Baldwinian processes. In other words, language may have started as a learned social behavior that proved so useful it fed back into our biology, shaping the genetic capacity to acquire and process language efficiently.
Pinker was right that language evolved biologically - there was some innate capacity for language encoded in human biology. Also, that language evolved gradually - we began making simple noises to indicate danger like other animals, before developing a more and more complex system resulting in the language we have today. Finally, he was right that language improves communication and coordination, which is crucial for social animals.
ReplyDeleteHowever, not every feature of language is purely optimized; we must consider other factors like cultural evolution and cognitive or developmental constraints. Also, language evolves faster than genes; sometimes new languages can spring up within a generation.
Hi Emma, your point about cultural evolution and specific languages evolving faster than genes is interesting and reminded me of something Professor Harnad discussed in class. I think these are a consequence of "lazy evolution" as he discussed, and are arguably more optimized than if language were coded in a less flexible way. "Lazy evolution" as applied here is probably the reason that language exists for us, yet we are completely flexible as to which one we learn -- it seems simpler to genetically code for some of the structural rules of language (like Universal Grammar suggests) than to code for an entire language's lexicon (plus structure would probably also still have to be coded anyways). You also mentioned cognitive or developmental constraints on language, but as Pinker and Bloom pointed out, there are common cases of language abilities and other measures of "intelligence" being seemingly uncorrelated.
DeleteDepending on what you mean by optimized, I think evolution optimized this trait fairly well!
8a. Pinker and Bloom make a strong case that language evolved through natural selection, but their view of how grammar developed feels a bit too clean. Evolution doesn’t just jump from no syntax to full syntax; there had to be messy, in-between stages. They don’t really explain what those proto-grammars might have looked like or how they worked. Still, I think their main idea makes sense: the structure of language seems too well-designed to be random, even if the path there was probably a lot slower and more complicated than they suggest.
ReplyDelete“Language is a complex biological adaptation to communicate information, shaped by natural selection for the transmission of knowledge.” Pinker and Bloom argue that language did not come from culture but evolved because it helped humans survive. Like the eye or the ear, it became part of our biology since it made sharing ideas and learning faster. Without it, people would have to learn everything on their own, which is impossible. I understand what the prof said about how no other species developed this kind of system. That makes language very special. I think what makes it unique is that it lets humans think and learn through symbols and words, not just by experience. It is more than communication; it is how we learn and think together.
ReplyDeleteHere, it seems to argue that language would have evolved in a Darwinian way through natural selection. It points out the many ways, like how transmitting knowledge and reducing ambiguity increase the odds of offspring surviving, and how it is very important to woo others with language, which works not so differently than mating rituals in other species. However, I wonder if we were to argue for language evolution, wouldn't it be better to think of it in a Baldwinian way? Something that was discovered as being beneficial, which then gets passed on to an offspring. Then, you eventually pass on the genes that make learning language easier. It could even lead to Chomsky's UG.
ReplyDeleteWhat captured my attention the most in Pinker and Bloom's argument is their insistence that grammar's arbitrariness doesn't undermine its adaptive value but it actually requires it. Their point about "parity in communications protocols" reframes what seems like design flaws as functional necessities - that arbitrary conventions are adaptive precisely because they're shared. This challenged my initial assumption that optimal design means eliminating arbitrariness. But I'm left wondering about their dismissal of spandrels. They acknowledge that not every feature needs individual adaptive explanation, yet they seem reluctant to admit language might contain genuine evolutionary byproducts. Could recursion, for instance, be exaptation from other cognitive systems rather than selected specifically for communication? Their argument works best when explaining broad design features, but struggles with explaining why these particular solutions emerged rather than alternatives.
ReplyDeletePinker and Bloom argue that human language can be explained by Darwinian natural selection. They suggest that grammar is an adaptation designed for communicating propositional structures. The authors explain that language must have been shaped by evolution based on its functionality and intricate design, comparing it to the adaptive complexity of organs like the vertebrate eye. Their view challenges Chomsky and Gould’s, which argues that language is an unintentional byproduct of other cognitive developments.
ReplyDeleteI think this debate is interesting in the context of LLMs like ChatGPT. Universal grammar proposes that an innate framework of grammatical rules exists, it is not learned. However, LLMs appear to acquire learnable grammar by finding correlations through unsupervised learning (from the Big Gulp). While ChatGPT’s sentences comply with Universal grammar, it is likely a reflection of the fact that the human authors of the Big Gulp inherently comply with Universal grammar, rather than the LLM having truly learned it.
“The key point that blunts the Gould and Lewontin critique of adaptationism is that natural selection is the only scientific explanation of adaptive complexity.”
ReplyDeletePinker and Bloom (1990) argue that language—particularly grammar—is an evolved product of natural selection. One of the main critiques of this view comes from evolutionary theorists Gould and Lewontin, who caution against assuming that all traits are direct products of adaptation, ultimately advocating a non-adaptationist stance that considers byproducts, constraints, or spandrels as possible evolutionary explanations. Pinker and Bloom counter that this critique loses strength when a trait exhibits functional, complex, and organized design—as language does. Therefore, when a system demonstrates a clear adaptive purpose and intricate design, natural selection becomes the most scientifically plausible explanation for its evolution.