🔀 How Metaphors Construct the World We Think We Know

See the corresponding Substack for more.

We're living inside a linguistic conspiracy so elegant that we can't see it. Every time you say "time is running out" or describe someone as "looking up to" another person, you're not just using colourful language—you're revealing the secret architecture of human consciousness itself.

This isn't another feel-good piece about the beauty of language. This is about power. The power of invisible structures to shape not just how we communicate, but how we think, what we can imagine, and ultimately, what kinds of realities we're capable of creating.

Diachronic data analysis supports and refines conceptual metaphor theory | PLOS Complex Systems

This is Heliox: Where Evidence Meets Empathy

Independent, moderated, timely, deep, gentle, clinical, global, and community conversations about things that matter.  Breathe Easy, we go deep and lightly surface the big ideas.

Thanks for listening today!

Four recurring narratives underlie every episode: boundary dissolution, adaptive complexity, embodied knowledge, and quantum-like uncertainty. These aren’t just philosophical musings but frameworks for understanding our modern world. 

We hope you continue exploring our other podcasts, responding to the content, and checking out our related articles on the Heliox Podcast on Substack. 

Transcript

0:25

Have you ever stopped to truly consider phrases like, time is running out, or, you know, looking up to somebody? We use them all the time, don't we? Almost without thinking. Exactly. But what if these seemingly simple expressions hold a kind of secret key to how our minds actually work? It's a really profound question, actually, because these aren't just clever figures of speech. They're more like insights, deep insights of the fundamental structures of how we think. Totally. So today on The Deep Dive, we're embarking on a really fascinating journey into the world of metaphor. But not just metaphor and poetry, right? No, not just as poetic flourish. We're talking about metaphor as a core, often hidden mechanism that shapes human thought and language itself. Okay. And our guide for this deep dive is some groundbreaking research. It's from a study called Diachronic Data Analysis Supports and Refines Conceptual Metaphor Theory. published in PLOS Complex Systems. The authors are Marie Teich, Wilmer Leal, and Juergen Jost. Yeah, and what makes this study so remarkable is its approach. It uses advanced computer-based data analysis to build what the authors call a solid and lasting bridge. A bridge between computational methods and cognitive linguistics. Which is basically the study of how language works in our minds. Exactly. Specifically, conceptual metaphor theory, or CMT. Okay, so our mission today is to kind of unpack how this research empirically tests these longstanding ideas about metaphor. It feels like we're finally getting a data-driven look, doesn't it? Yeah. At how meaning has really emerged and evolved in English over centuries. It does. It's almost like a shortcut to seeing how language and thought are truly, like, braided together, revealing these surprising structures just beneath the words we use every day. So, to start, we probably need to grasp the core idea of conceptual metaphor theory, CMT. Yeah. Which was, you know, widely popularized by George Lakoff and Mark Johnson. Their big book was "Metaphors We Live By." Right, a classic. And their central argument was that metaphors are way more than just rhetorical devices. They're not just for fancy speeches or poems. Okay. Instead, they propose that metaphors are central cognitive linking mechanisms. Cognitive linking mechanisms. like bridges in the mind. Sort of, yeah. Yeah. They connect our concrete, embodied experiences, you know, things we can physically touch, see, feel, to more abstract linguistic domains. Okay, give me an example. Well, think about love as a journey. The journey part comes straight from our physical experience of moving through space. That makes sense. And that concept, that structure of a journey, gets systematically transferred. It becomes a fundamental part of how we understand something abstract, like love beginnings, obstacles, destinations. That's a powerful claim. Love isn't literally a journey, but we think about it that way. Exactly. But despite how influential CMT has been, it's also faced, well, constant criticism over the years. Really? Why? The main issue was this perceived lack of systematic empirical research, research based on actual linguistic data. Ah, so it's mostly based on intuition. A lot of it relied heavily on researchers' linguistic intuition, which is valuable, don't get me wrong, but it isn't always easy to scale up or verify objectively. And this is precisely where this new research steps in, right? to tackle that empirical gap. Exactly. They tackle it head-on using something called diachronic data analysis. Diachronic meaning across time. Precisely. So they analyze language changes across the long history of English. You can kind of think of it like this. Okay. The study treats this historical language data as archaeological artifacts of past metaphorical events. Oh, I like that. Like digging up old tools. Yeah. It's similar to how an archaeologist studies ancient ruins or tools to understand how people lived and thought thousands of years ago. by looking at the linguistic fossils left behind. The words and phrases. They can indirectly approach the events themselves. Those moments when metaphors first formed and started shaping meaning. That's a great analogy. And it's important they're not saying computers understand meaning like humans do, are they? No, no, absolutely not. It's more that they're using sophisticated mathematical structures to enable this huge corpus-based statistical analysis, but framed within CMT's ideas. It's about finding patterns in massive amounts of language data. So what was the big reveal? What did this deep dive into the data confirm? Well, the study's core finding really confirms that fundamental assumption in CMT, metaphors are indeed enduring linguistic and cognitive structures. Not just fleeting rhetorical figures. Exactly. Not just things that pop up and disappear. How did they show that? They saw a significantly higher probability of transferring words along already established Yeah. mappings. Oh, so once a path is there, it gets used again and again. Precisely. This suggests what they call persistent metaphoric mapping structures. So for you, the listener, what this really means is that the language we use, especially those deeply ingrained metaphors, it actively shapes our cognitive system. It builds the roads in our minds almost. Kind of. It creates these built in pathways like super highways for new ideas, making certain new meanings much more likely to emerge along those existing conceptual routes. Okay. So how did they map this out? You mentioned networks. Right. To really unravel this, the researchers thought about metaphor formation as a network, like a vast web. Okay. Imagine individual topics or themes like time or emotion as the nodes, the points in the web. Got it. And the connections between them, the edges, represent these metaphorical transfers, moving from a source topic to a target topic. And where did they get the data for these connections? They used an impressive data set from the Mapping Metaphor Project. It's this huge collection of etymological transfers, word history, showing meaning shifts caused by metaphorical mappings. Wow. Across 415 thematic categories from the Historical Thesaurus of English. So a really rich data set. And what did this network look like? Was it just a big jumble? No. And that's what's really striking. They found this metaphor network is absolutely not random. Okay. So there's a structure. A significant hidden structure. Yes. Most categories, most topics show low activity. But there's this fascinating heavy tail, a smaller number of highly active categories that really drive the whole network's dynamics. Interesting. And what was the biggest structural finding? The most pronounced property they identified is something they call the systematic separation of categories into two distinct elements. anti-communities. Anti-communities. That sounds adversarial. Huh. Maybe a little. Yeah. But think of it less as opposition and more like two distinct clubs in the network. Okay. Where members of one club mostly interact with members of the other club rather than their own. Ah. Okay. So they connect across groups more than within groups. Exactly. One of these anti-communities, anti-community one, is what they labeled the concrete group. Concrete. like physical stuff. Mostly, yeah. Things that are spatial, mechanical, physically perceptible, bodily experienceable domains, as they put it. Example. Categories like food and eating, shape, relative position, things like that. And what's interesting is these concrete topics serve as both strong source domains and target domains for metaphors. They give and receive meaning. Okay, so concrete stuff, gives and receives. What's the other club? That's anti-community too, the abstract group. This is mainly temporal, emotional, and social topics, like bad, emotional suffering, literature, time. Right. And crucially, these categories primarily act as target domains. They receive metaphorical meaning from the concrete group, but they rarely serve as sources themselves. So the flow is mainly from concrete to abstract. Right. The overwhelming majority of connections are between these two classes, yeah. Connecting concrete ideas to abstract ones. Much less connection within the abstract group or within the concrete group comparatively. That makes intuitive sense, actually. That lines up with a lot of the classic CMT examples. It does. And this structure reveals two dominant metaphorical processes they identified. Yeah. Okay, what are they?

First, the most common we just talked about: 8:45

Maffings, from the concrete anti-community to the abstract one. That's the classic pathway. Right. And the second... But here's where it gets really insightful. The second major process involves mappings within the concrete anti-community itself. Wait, concrete concepts mapping onto other concrete concepts? Exactly. This is a less explored area, and the study suggests it shows the continuing redetermination and relinking of concrete themes. That finding the frequency of metaphors within the concrete group That feels like it challenges simpler ideas about embodiment, doesn't it? It really does. It's not just about taking our bodily experience and slapping it onto abstract ideas. It points to something far more, well, multi-directional. An interweaving of imagination and experience, as they put it. So creating meaning might involve constantly, like, re-experiencing the concrete. That's the idea. Metaphor might help us continuously reorganize and reconnect our bodily perception itself. Wow. So it implies metaphor isn't just based on embodiment. It might actually be an instrument for our constantly renovating embodiment. Renovating embodiment. I like that. Like we're constantly updating how we understand our physical world through these conceptual links. Precisely. How we reform and deepen our understanding of our own physical presence and how it relates to thought. Okay. So we've got these deep persistent structures, these anti-communities, these flows of meaning. But how do metaphors actually form and become so persistent? Right, the dynamics. Yeah. And the study had some really surprising things to say here, didn't it? Challenging some core CMT assumptions, like about space. Yes, exactly. A central claim of CMT has always been that spatial thinking, you know, up, down, front, back, near, far, is this fundamental source for all concrete structures. Like a universal platform. Seems logical. Was that confirmed? Well, interestingly, no. The new research found something different. They state, none of the analyses conducted on the mapping metaphor data showed a central, highlighted role of the spatial and directional categories in the metaphor network. Really? So space isn't the kingpin? It seems not. They observed that spatial domains show a similar behavior as other concrete domains, not some special top-tier rank. Huh. So our reliance on spatial metaphors might be just one major force among others, not the underlying source of everything. That seems to be what the data suggests. It's still important, just not uniquely fundamental in the way CMT proposed. That's genuinely surprising. Okay, now let's dig into the mechanism of metaphor formation. How do these connections actually get me? They did something called motif analysis. Yeah, motif analysis. It's basically looking for small recurring patterns of connections in the network, little subgraphs. And what did they find? What they found was almost an absence of something called closed triangles. Closed triangles, meaning? Meaning transitive connections. Yeah. Where if A maps metaphorically to B and B maps to C, you'd expect A to also map to C directly, that ABC structure resulting in AC. Okay. And those were missing? Largely absent, particularly between the two anticommunities, the concrete and abstract groups. So if argument maps to war, B, and war maps to destruction, C, it doesn't automatically mean argument maps directly to destruction. Exactly. The path doesn't automatically close like that in the data. So what does that imply about how metaphors form? Well, this is a really significant revelation. It suggests that metaphor is not primarily triggered by some pre-existing structural similarity or analogy between two categories. Wait, so we don't see that love is like a journey and then make the metaphor? The study suggests it might be the other way around. It's likely, they propose, a tension between two categories that invites metaphors and thereby reveals structural similarities. Wow. Okay, so the metaphor doesn't just describe a similarity. It creates it, or at least brings it into focus. It highlights it. That's a huge distinction. It's like the metaphor itself makes us see the connection rather than just reporting one that was already obvious. It's a profound shift in understanding, yes. The metaphor actively structures our perception. Mind blown. Okay, what else did they find about the mechanism? Something else really interesting was the high presence of symmetric metaphorical links. Symmetric, meaning? Meaning if a source-to-target mapping is likely, like concrete A, abstract B, then a target-to-source mapping, abstract B, concrete A, also becomes more likely. So it's not always a one-way street from concrete to abstract. The abstract can influence the concrete. Exactly. It gives a kind of feedback effect. The target domain can actually reconceptualize the source domain. Huh. Can you give an example? It's complex, but imagine how thinking about arguments as war might, over time, make us perceive actual war in terms of positions, defenses, attacks, concepts, reinforced by the argument metaphor it's subtle okay I think I get it so metaphor is not a purely unidirectional process it's more dynamic a two-way interaction very much so both concepts can evolve in relation to each other through that metaphorical link which brings us back to one of their strongest confirmations The sheer persistence of these mappings. Right. They found that significantly higher probability of transferring words along already established mappings. This is the first real hard data for that. They claim it provides the first systematic and data-based evidence for persistent metaphoric mapping structures. So once that conceptual road is built, it becomes a major highway for future meaning making. Indeed. Once a metaphorical connection is established, it doesn't just fade. It becomes a strong, enduring pathway for more linguistic development, more meaning creation. It really reinforces this idea that metaphors are deeply embedded structures in our cognition. Okay, so beyond just finding these structures, did the study offer a new way to map meaning itself? Yes, definitely. They identified unique semantic roles for categories within this metaphor network. Semantic roles. Yeah, using a technique called Hierarchical Cluster Analysis, or HCA. It basically groups categories together based on how they behave in the network, what kinds of metaphorical connections they send out and receive. It reveals how different topics play similar roles in the overall structure of metaphorical meaning. And how does this compare to how we usually map meaning, like in AI? Ah, well that's what's truly captivating here. This metaphor-based semantic structure is unique. How so? It has low correlation and little mutual information with semantic distances you get from typical word embeddings. Okay, hang on. Word embeddings, those are the computer things, right? Right. Good point. I think of word embeddings as how computers generally understand words now. They turn words into numbers, vectors, based on the context they usually appear in. So words that show up near similar words are considered close in meaning? Basically, yes. It's based on co-occurrence. It's how search engines or chatbots process text. Got it. So if I'm hearing you right, this metaphor map is fundamentally different from that. It's like our current AI is missing this whole figurative layer. That's a great way to put it. This study suggests we're potentially mapping out that missing piece. This new structure encodes a semantic position that goes beyond the semantic information contained in word embeddings. It captures figurative meaning. Exactly. It encompasses much richer figurative and conceptual semantic similarities. Yeah. It's like a whole different dimension of meaning that co-occurrence alone doesn't capture. That's huge. Did they give examples of this information? figurative closeness? They did. Some really fascinating ones, like the connection between fluidity and life. Oh, yeah. The study found this deeply rooted conceptual proximity between dryness and absence of life on one side. Oh, okay. Makes sense. And then flowing and floating, wetness and humidity, grouped closely with life on the other side. Wow. It shows this parallel semantic structure, even though there's no direct literal overlap in the topics themselves. That's incredible. So it's not just that water literally sustains life, but our very concept of life is tied to flow and wetness at this deep metaphorical level. That's what the structure suggests. It really changes how you might think about, say, environmental discourse or just how we perceive a desert versus a river. Totally. What about emotions? They also found interesting emotion groupings. Certain strongly moving emotions, they call them, like excitement, anger, and emotional suffering. These group together with categories like intelligibility, like understanding, and also moral evil. Anger and suffering linked to evil and understanding. That's complex. It is. But other emotions like fear or esteem were not in this group. So it's not just all strong emotions are the same. No, it maps out a much more nuanced figurative landscape. It shows how concepts like anger and suffering are metaphorically linked in our conceptual system to how we grasp badness or even clarity, maybe through intensity. But fear operates differently in that cognitive space. Wow. Okay. Any other cool examples?

Get this: 17:54

different types of movement weren't all lumped together either. No. Instead, they formed distinct motifs reflecting unique conceptual links. For instance, continuation grouped with intention and planning. That makes sense. Planning involves continuing in action. Right. But then repetition was grouped with sky and celestial sphere. Sky. Celestial sphere. Why repetition? Maybe hinting at the cyclical nature of the sun and moon. or the seemingly endless expanse implying repetition. It's suggestive, right? Yeah, definitely more abstract. And then progressive movement moving forward was surprisingly connected to impulse and impact, like a physical force. Yeah. And even deity. Deity. Like God. linked to moving forward. It seems so in this conceptual map. Perhaps linking divine will or powerful forces with the idea of progress or forward motion. These aren't intuitive links based on topic similarity alone. No, they're definitely based on some deeper figurative logic. So what does this new map mean for us, for you listening? Well, the potential is huge for understanding language and thought, obviously. But practically, it could really enrich natural language processing NLP, adding this crucial conceptual and figurative dimension to how computers process and understand language, moving beyond just word co-occurrence. Yeah. Making AI a bit more human-like and understanding. Potentially, yeah. And it could be a really powerful tool in discourse analysis. Imagine analyzing, say, a political debate. You could use this map to classify shifts in how a topic is framed. Is the shift nuanced, moving to a conceptually nearby category on this metaphor map? Or is it profound, jumping to a completely unrelated conceptual area? That could give incredible insight into rhetoric and... and persuasion, seeing if someone's just slightly reframing or making a radical conceptual leap. Exactly. It offers a data-driven way to analyze those kinds of communication strategies. Okay, so let's try and wrap this up. To recap this deep dive, we've seen that metaphors are not just, you know, pretty words. Not at all. They're fundamental, persistent cognitive structures. Right. They actively shape our language and our thought. And this study gives us powerful empirical proof. It really does. It confirms and refines these long-held theories from CMT, but with a new data-driven perspective. Yeah. A whole new map of meaning, almost. It truly makes you think that language isn't just a tool for expressing thoughts we already have. No, it seems much more dynamic, like an ever-changing system that is constantly creating and reorganizing meaning. Even at that really fundamental embodied level we talked about. Especially there, perhaps. So here's a final thought to leave you with. If a metaphor is constantly reorganizing the relations between concrete domains, like the study suggests, and renovating our embodiment, what does that really suggest about the nature of Hume experience itself? Does it imply that our perception of the world, maybe even our own bodies, is far more fluid, far more imaginatively constructed than we usually give it credit for? It's a fascinating question. How much of our reality is built through these hidden linguistic structures? Something to definitely ponder as you go about your day. Start noticing the metaphors shaping your own thoughts.