In a Part I of this post, I summarized what I thought were the most interesting points in Steven Pinker’s The Language Instinct. Now, I’d like to continue on to his, in some ways, more impressive book, The Stuff of Thought: Language as a Window into Human Nature (link on Amazon.com). Since they are both related to language, how might I characterize the relationship between these two books? Here’s what I would say: The Language Instinct explores a tool that we all use and have some appreciation for, language, and shows how exquisite and breathtakingly-powerful it is, as if we were in the 1920’s, driving around in Model T Fords, only to find that underneath the exterior we were all driving 1990’s Lamborghinis. The Stuff of Thought examines the structure and features of these Lamborghinis and speculates on what they reveal about the drivers/creaters of them. I found the first book to be more profound, and the second book to be more fascinating.
The primary theme in The Stuff of Thought is that we have developed and shaped language to serve our functional aims in communicating with others about the world around us (Pinker’s phrase is being able to communicate ‘who did what to whom’), thus analyzing language usage reveals insights about these functional aims. Although the book discusses Standard American English, many of the insights are similar to, or have parallels in, many other languages. Below I will list four summary points of what I found to be most interesting:
1. We refer to physical space and the objects within it using prepositions based on the functional possibilities of the spaces and objects. When we refer to where things are in space, we use prepositions such as ‘in’, ‘on’, ‘under’, ‘by’, ‘next to’, ‘on top of’, ‘near’, ‘far’, etc. But what governs the fine distinctions of when we use one preposition or another? Consider the two pictures below. Would you agree that in the left picture, we would say that the ball is on the hand but that in the right picture the ball is in the hand?
Why is it that ‘on’ seems correct in one case but ‘in’ seems correct in the other? If you would venture that it is because ‘in’ requires some sort of concave structure or container, then that appears to be right, but then how much concavity is required to transition from ‘on’ to ‘in’? Picture the palm on the left slowly curling around the marble, and you can imagine when you might switch the preposition from ‘on’ to in’. Pinker suggests that this transition occurs when the hand can functionally control the object, move it around securely, etc. This is an example of a functionally-based hypothesis of language development (some aspects of which can be tested in the lab, by the way). This might explain why we would ask a young boy why he has a caterpillar ‘on his hand’ if he is just watching it passively, but ask him why he has a caterpillar ‘in his hand’ (with the caterpillar and his hand in the same position), if he is walking towards the house with it.
Another interesting way we talk about space is how we talk about discreet objects in our environment. Although physical objects have three dimensions, sometimes they ‘lose’ dimensions when we talk about them. For example, when we refer to large bodies of water or areas of land, we usually talk about them as if they were flat 2-D surfaces. This explains why we can use the preposition ‘under’ in “he nearly drowned under the water”, and use the words ‘underwater’ or ‘underground’, though they refer to ‘under the water line’ and ‘under the ground level’, which are their relevant 2-D surfaces. Another example is when we talk about a wire, which we often reduce to a 1-D surface. When there are instructions to cut the middle of a wire, people don’t wonder if they should try to split it lengthwise. Of course we can talk about these less-referenced dimensions if we need to (“try to choose a thicker wire”, “the lake is deep enough to hide the Loch Ness Monster”). In some cases when we may even reduce objects to zero-D blobs when the geometric proportions are not relevant. For example if Jim is riding in the passenger seat of a car with his head and one arm sticking out of the window, we still say that ‘Jim is in the car’, because for our functional purposes, we are describing him as being transported in the vehicle–he loses all dimensions and becomes a blob(sorry, Jim) . We’ll see that this human-action centered way of describing the world carries over in the realm of time as well.
2. We refer to time by dividing it up into what can and can’t be changed and also pay particular attention to when activities start and stop. In most languages we divide up time into three main categories, the past, the present, and the future. Of course we can be more specific by adding adjectives or time references, but the basic divisions are the past, present, and future. The past consists of things that occurred and cannot be changed, starting from the beginning of the universe until just a moment ago. The present is what is happening ‘now’, a very short period of time, suggested to last about three seconds (the time to make a quick unrehearsed decision, the decay of short-term memory, a typical line of poetry or musical motif, etc.). It goes without saying that the functional importance of ‘now’ offers sufficient reason to give such a short period of time its own tense. The future starts a moment from now and lasts until the end of the universe. The interesting thing about how we talk about the future is that it seems to incorporate an implicit theory of free will. In English, as well as other languages, we often use the same grammatical constructions for certain statements about the future and our own future intentions. Thus, we may say “The comet will be visible at 1:45 pm tomorrow” or “I will be in my opera box for the opening curtain”. Depending on when in the past it was uttered, “The Miami Heat will advance to the 2013 NBA Finals” is a logical statement of fact or the confident boast of an ardent fan.
There is one other facet of time that we can focus on in English, should we choose, which is when things begin and end (referred to as aspect), and we can specify the aspect whether we are talking about the past, present, or future. For example, we can describe beginnings of actions or states even without using the words ‘begin’ or ‘start’, such as when we say “I dove into my work after lunch”, “He entered the room”, or “Please lift the lid for me”, all of which have clear starting points but indefinite end points. Conversely, there are ways we focus on the endings, while being non-specific about the beginnings: “He reconciled with his wife”, “The chicken just crossed the road”, or “She’ll decide on a major by the Fall”. There are also ways to describe activities without referencing either beginnings or endings, e.g. “We were eating dinner when you called”. Now, it may be obvious why we divide up time into past, present, and future, but why do we have such precise methods available to describe aspect? Again, it may be for functional purposes, where knowing when things began and end may help establish causal linkages, and make clear what activities have started and what states have changed (there’s a big difference between ‘he put out the fire’ vs. ‘he is putting out the fire’!). To further illustrate this, look at this sentence I’ve constructed:
“When I told him that his grandmother’s Tiffany-esque lamp was likely worthless, his eyes started to well up and he told me about his 10-year struggle with gambling…”
You can see that even though I’m describing an event in the past, elaborating on the aspect allowed me to convey causal sequences with the result of painting a rather pitiable picture of this fellow’s emotional state and the reasons behind it.
3. We use a large number of metaphors implicitly in English, though many have lost their analogical power. Many of us have had occasion to reflect on how prevalent metaphors are in the way think and talk. Memorable speeches use clever metaphors to capture our attention and illustrate a point. And in the high art of ‘literary fiction’, generating an unusual metaphor by pointing out some parallel between jarringly disparate images that evokes some subtle insight, is a mark of a writer’s power. At the beginning of chapter 5 of the book, Pinker analyzes the introduction to the U.S. Declaration of Independence (“When in the course of human events…”) and dazzlingly draws out more than half a dozen implicit metaphors, all from that famous single sentence.
One particular metaphor is particularly deeply embedded, that of thinking of time in spatial terms. One theory of the prevalence of this TIME IS SPACE metaphor is that space is concretely experienced by our senses, while time is more abstract, and thus we naturally think of the latter along the lines of the former. The dominant time metaphor is that of time as a landscape rolling past us, as if standing in a river, thus, “summer is here”, “the weekend went by quickly”, “she has a bright future ahead of her”. This metaphor is so deeply embedded that we usually don’t even notice it. This metaphor is not exclusive, however, as there is a competing metaphor of time as a stationary landscape that we are travelling through, thus “we’re approaching our anniversary”, “I’ve reached middle-age”, “after me, no one managed to get past level 14”. Note that we can switch between these two metaphors easily, but in some cases without sufficient context, ambiguous statements are possible, such as “Let’s move the meeting ahead two weeks”—is that two weeks earlier towards us or two weeks later ahead of where we are? (Note: I do wonder if we could combine the two metaphors into a single one of us walking as on a treadmill, or in a river, with both ourselves and the landscape moving toward each other, and each of the two metaphors reflecting a different perspective.)
Other common metaphors are LOVE IS A JOURNEY, which is embedded in expressions like “We’ve come a long way together”, “Our relationship has stalled”, or “Our therapist says we’re making progress”, and ARGUMENT IS WAR, yielding “He attacked my position”, “She defended herself well”. These metaphors are so deeply embedded and fecund that people can create novel expressions that build upon these that will be readily grasped by the audience—even if they are unusual or silly, for example: “They agreed to take a break from their relationship, but he found an oasis and threw away his canteen” or “She totally out-prepared him in the debate—while he was sharpening knives, she was enriching uranium”.
Two more points about metaphors. One is that many of them have been incorporated into common utterances so long that they have turned into figures of speech. Studies have been done measuring reaction times to comprehend various metaphor-based sentences, with subjects either primed or not towards the metaphor, and indeed they find that comprehending fresh metaphors does respond to priming, but well-established metaphors are processed just as quickly either way, and thus can be thought of as dead metaphors. The other point is that although metaphors are deeply embedded in our speech community and draw upon our minds’ ability to think analogically, it does not follow that metaphors largely determine the way we think. There can be strong influences in some cases, no doubt, such as when we conjure up images of ‘trickle down economics’, ‘workers as cogs in a machine’, ‘government as big brother’, ‘environment as a victim’, etc. However, besides the fact just pointed out that over-used metaphors can effectively ‘die’ with respect to their analogical power, we see that people still judge the aptness of metaphors (“no, that’s not a good parallel…”), we often mix metaphors to suit our purposes, and new metaphors are being generated constantly by comedians*, little children (“can you ladder me up the stairs?’), authors, and everyone else. Our amazing facility to grasp these metaphors and see the analogical connections behind them hopefully keeps us mentally flexible enough to be critical of them when necessary.
4. We have incorporated into our language an intuitive pre-Newtonian physics and an implicit theory of free will. Let’s say we observe a boulder that rolls down a hill (from a rockslide, or perhaps someone pushed it) and it hits a fence and knocks the fence over. The way use language to describe this incorporates an intuitive physics. It closely resembles an ‘impetus’, or force-dynamic paradigm, where we envision that the rock is initially endowed with a certain impetus that set it in motion, and then when it strikes the fence, it transfers some of that impetus to the fence, and if that is stronger than the inherent impetus of the fence to stay still, will cause the fence to give way. Thus, we may use phrases such as the following, which reflect this force-dynamic view:
- The rock’s force overcame the strength of the fence
- The fence could not resist the force of the rock it was overwhelmed
- The fence was not designed to withstand such a large impact
- I knew the fence wouldn’t hold up to such an overpowering force
If we were describing things based on a Newtonian point of view, we would talk about a collision between two objects and the effects of the impact on each, rather than the narrative framework of an actor acting on a resistor, and either being successful or not. However, when we realize this is developed with a pre-Newtonian understanding of how we see cause and effect in nature, this makes sense. We also talk about helping or hindering the active agent or the resistive object as follows:
- The wind helped push the rock along, and enabled it to break the fence
- The rock rolled down despite the grass slowing it down
- Steve braced himself against the fence, preventing the rock from knocking it down
Of course most descriptions of causation that we talk about involve living agents, and in this, we tend to originate causes in the black box of an agent’s mind. In physics we don’t describe things as uncaused causes, but we do this all the time when talking about everyday life. If we describe Joe hitting Steve (maybe for holding up the fence), we don’t talk about low-level physical causal events, such as muscle motors and brain signals, nor do we usually talk about higher-level causes, such as the elite’s exploitation of the working-class, we simply say that Joe hit Steve, and then judge Joe as an autonomous actor. Note that in modern times we do sometimes talk about these other levels of causality, at a lower level when we talk about someone acting due to psychopathy, and at a higher level when we justify some act as a response to a societal pressure—however, this is not the norm in most conversations. The same is true for other living agents, we say that a wasp stung us, not that the hive instinct, as executed through a wasp, stung us. In contrast to active agents, we don’t blame a tree when a branch falls on us, or a mountain when we are caught in a rockslide, rather we blame luck or ourselves for putting ourselves in that situation. Living agents with autonomous movement are a source of uncaused causes, at the level of explanation we use.
The force-dynamic view may also influence how we distinguish what level of human causality is considered praiseworthy or blameworthy. We distinguish actions that supposedly directly cause a result vs. actions that merely enable that result. Let’s say that Joe throws Steve on to the street, into the path of an oncoming bus. The bus driver, showing extreme prejudice against jaywalkers, can easily stop in time, but chooses not to do so and runs over Steve. Our intuition as we consider this situation is that we say that Joe caused Steve’s death, but that the bus driver only failed to prevent it. Even if Steve and the bus driver both testify in court that they each wanted Steve dead, we would assign the majority of the blame to Steve. This is despite the fact that they were both in similar circumstances in that each had a choice of actions that could lead to Steve’s harm, and both took it. The way we tend to view it is that Joe not only wanted to kill Steve, but that he initiated a causal chain of events that led to his probable death. We like to keep the causal chain close. For example, if Mary were to talk Joe out of his act, we might say she prevented Steve’s death, but if Daniel stopped Mary before she could stop Joe, we wouldn’t normally say that Daniel was ‘responsible’ for killing Steve. When it comes to doing harm, how directly we cause it makes a lot of difference, and it is illustrated in the famous Trolley Problem, which has much discussed in philosophy, legal studies, and psychology:
Imagine that you are aboard a train that is headed towards five workers who will be caught unawares and run over by the train. There is an alternate track, however, and you can throw the lever to switch to the other track, which has one person on it. Thus, if you do nothing, five people will be killed, and if you switch tracks, one person will be killed. What do you do?
Now consider this variation: You are on a footbridge above the train track and you see the train coming on towards the five workers. Beside you is a large man, whom, if you pushed onto the track, would certainly stop the train and spare the five workers. Do you do this?
The majority of people will answer that in the first dilemma they will switch tracks, because one death is better than five deaths, but in the second dilemma most would not push the large man over. Yet in both cases, by your action you could cause one person to die instead of five. Think about your own answers to these dilemmas and what is driving your intuitions. There has been much written about this, and you could offer many justifications for the different choice in the two cases, but in short, we seem to make a distinction between directly causing harm by initiating a force (pushing) versus allowing an event occur (letting), with a stronger sanction on the former (see here for more commentary on the Trolley Problem). But upon closer inspection you can see that it is not the pushing that kills the man, gravity and the ground do the work—but our minds tightly group together the causal events intended by the free will of an agent, initiated by a physical movement, and leading directly to the physical result.
The above convey some of the major discussion topics in The Stuff of Thought. Other topics that are covered include: verb sub-classes with fine-grained shades of meaning (which we apply almost subconsciously), a discussion of swear words and taboo words, a philosophic perspective on assigning proper names, and a brief introduction to pragmatics (that area of linguistics which deals with conveying meaning outside of strict grammatical sentence construction). Reading the book was simply thrilling because every chapter pointed out numerous quirks in our everyday language use that we hardly even notice, and then gave a compelling explanation for them, so it simultaneously increased my awe at the complexities of language and generated many insights into how our minds think about the world, ourselves, and society around us.
Besides reading the book (which of course I highly recommend), you can listen to Steven Pinker’s hour-long talk at Google in 2007:
For a 20-minute version, here is Pinker’s 2007 TED Talk: