↵ Back to class homepage

Figuring out the truth-conditions for statements about specific people or things is relatively simple. (Relatively... if you take a class on semantics you will learn lots more complications I'm ignoring here.) For example, Haile is a doctor is true if and only if the person named Haile in the relevant discourse context is a doctor. (As we saw in the previous module, there are also presuppositions associated with this sentence, e.g. the presupposition that there is someone named Haile, and that there is only one person named Haile in the context of our conversation. But it's not clear that those are part of the truth-conditions of this utterance.)

But what about a statement like "Boys are good at STEM"? What conditions are necessary for this statement to be true? Is it true if only three boys in the whole world are good at STEM? Is it only true if every boy in the world is good at STEM? Or what?

Statements like these are called generic statements, or generics for short. Rather than stating something about a specific entity (a person, thing, etc.), they instead describe a generalization about some whole category of entities. And it's not actually clear what these statements mean. We know this both from anecdotal experience (there are many real-world examples of different people interpreting a generic statement in different ways), and from controlled experiments, such as one by Cimpian and colleagues (2010), the paper from which most of the information and concepts in this module come from.

An example of a generic statement interpreted in different ways

We can find a nice example of this in a discussion of putative Eastern/Western cultural differences on Language Log. The article which this blog post is critiquing contains a generic statement, "Americans usually see individuals; Chinese and other Asians see contexts."

The statement "Americans usually see individuals" is extra-complicated here because the word usually makes it ambiguous: does it mean that each American usually [but not always] sees individuals, or does it mean that most [but not all] Americans see individuals? The "not all" and "not always" bits are themselves quantity implicatures, not necessarily part of the literal meaning of the statement. The "each" and "most" bits are the parts that are related to the generic statement. But since this statement is complicated by the ambiguity, let's instead focus on the second statement, "Chinese people see contexts". (Note that I have paraphrased this a bit from the original, but I think this is justifiable; "Chinese and other Asians see contexts" semantically entails "Chinese see contexts", and it's clear that Chinese here means Chinese people.)

In the context of this discussion, "seeing context" is a certain way of behaving in a particular sort of psychology experiment. In the experiments that the writer of this statement is describing (and that the rest of the blog post is criticizing), people see a video of fish swimming around in a tank, where one fish is more salient (bigger, or a different color, or whatever) than the others, and the person watching the video is asked to describe the scene. Some people describe what the salient fish is doing; other people describe other aspects of the video, like the background or this other fish. This latter group of people are the ones who supposedly "see context", according to this experiment.

The big question for us, though, is: what does "Chinese people see context" mean? Does it mean all Chinese people see context? Most Chinese people see context? Chinese people see context more than some other group of people does? Seeing context is an important and defining feature of being a Chinese person? Or what?

It is on the basis of statements like this that people make further generalizations like "Westerners are individualist and Asians are collectivist" (or "Western societies are individualist and Asian societies are collectivist"). It is certainly questionable whether the results of that fish-describing experiment license these more general conclusions (let's not mince words, it's more than just "questionable"; it's obviously false). But that's not the relevant point here. The important point here is, those generic statements raise the same kinds of questions: does "Westerners are individualist" mean all of them are, most of them are, or what?

Further down in the comments on this same blog post, the author of the blog post writes, "The question is whether it's in some sense scientifically established that all Asians are collectivist and all Westerners are individualist" (my emphasis added). But is that the question? If you ask this question (and if you go on to say, e.g., "I have proved that it's not true that 'Asians are collectivist', because here I found an example of an Asian who is not collectivist"), you are interpreting the generic statement "Asians are collectivist" as meaning "All Asians are collectivist.

We can find many examples, though, of generic statements that people consider "true" even though a corresponding "all" statement is certainly not true. For example: "Ducks lay eggs". People widely consider this to be true. But it certainly is not the case that all ducks lay eggs. Male ducks don't lay eggs; baby ducks don't lay eggs; ducks that are too sick, old, or injured to lay eggs don't lay eggs. So, most ducks don't lay eggs. Similarly, Cimpian and colleagues point out that people usually judge the sentence "Mosquitos carry the West Nile virus" as true, even though only about 1% of mosquitos actually carry it.

The same applies for the present example. It is likely that many people would judge sentences like "Chinese people see context" or "Asians are collectivist" as true, even based on just a few supporting examples. But on the other hand, people might then take those same sentences to mean that all Chinese people see context or that all Asians are collectivist.

The "asymmetry" of generics

This example demonstrates a common "asymmetry" in how generics are interpreted: we don't need much evidence to accept a generic as true, but when we have to actually interpret it we think it refers to nearly all of the things in the category. Hence the title of Cimpian and colleagues' paper: "Generic statements require little evidence for acceptance but have powerful implications".

In their research, they gave people generic statements about made-up things, like "Lorches have purple feathers" (there's no such thing as a lorch!). Along with those statements, they provided statistical information, such as "30% of lorches have purple feathers", and then asked them if the generic statement ("Lorches have purple feathers") was true. They found that people tend to accept these statements as true even when the actual percentage of lorches with purple feathers is not very high, e.g. around 50%. (And people were even more willing to accept these if the feature being described is dangerous, distinctive, or otherwise striking and attention-grabbing; recall that, as we saw above, people believe "Mosquitos carry West Nile virus" is true even though only 1% of mosquitos do!) Thus the first half of this asymmetry: "generic statements require little evidence for acceptance". We are often willing to accept a generic statement (like "Boys are good at STEM", "Asians are collectivist", or "Lorches have purple feathers") based on only a few examples.

In this same experiment, the researchers also had people interpret what they think the generic sentence means—i.e., they asked, "What percentage of lorches do you think have purple feathers?" People reliably estimated a much higher percentage than what they needed to accept the statement as true. For example, people might accept the statement "Lorches have purple feathers" as true when just 70% of lorches have purple feathers; but if you tell people "Lorches have purple feathers" and then ask them what percentage of lorches they think have purple feathers, they will guess that 95% do! Thus the second half of this asymmetry: "generic statements have powerful implications". When we hear a generic statement like "Boys are good at STEM", "Asians are collectivist", or "Lorches have purple feathers", we often interpret it as saying something about almost all boys/Asians/lorches.

Real-world disagreements over generic statements

These properties of generic statements make them a common source of controversy, especially in political discourse. They can be used to suggest very broad generalizations, but they don't require very much evidence, and thus they can be pretty easily defended if someone criticizes them. Like we saw in our initial example above: if someone interprets "Asians are collectivist" as meaning that all Asians are collectivist and then brings up some compelling evidence that not all Asians are collectivist, someone else could respond by saying "Well, I never said that all of them are!"

For example, in debates over responses to the COVID-19 pandemic, we have all heard many statements along the lines of "Travel restrictions are an effective way to stop the spread of the virus" or "Vaccines work!" Generic statements like these are difficult to argue against because the meaning of generics is fuzzy. (I am not raising this example to suggest that vaccines are bad—I myself am vaccinated and I encourage everyone to get vaccinated if they are able—but merely to point out that generic statements come from everywhere on the political spectrum and not just from the people we think are manipulative bad actors. It's easy to use pragmatics concepts to criticize the language of people we don't like or don't agree with, but it's important to apply the same scrutiny to our own "team" as well.)

For another example, let's consider one of the first and most-remembered of many controversies in Donald Trump's 2015-2016 United States presidential campaign. Describing people who immigrate from Mexico to the US (it's not clear to me if he was talking about all Mexican immigrants or just undocumented Mexican immigrants), Trump said, "They're bringing drugs. They're bringing crime. They're rapists. And some, I assume, are good people." These remarks were widely criticized, as they should be, but also prompted widespread debate over what they actually meant. Many people incorrectly claimed that Trump said all undocumented Mexican immigrants (or even "all Mexicans") are rapists. But he didn't actually say that; he said a generic statement. Thus, in response to those criticisms, Trump supporters often replied, "He didn't say they all are!". And in fact they were right; the last little bit of the quote makes that explicit (as long as we adopt the extra assumption that "good people" are not rapists; this seems like a very reasonable assumption to me). My point here is not that what Trump said was good; it certainly was not, for other reasons (e.g., due to the Gricean principle of relevance, the fact that these characteristics are the ones he thought important to highlight certainly says something, and there is no doubt that these and similar remarks incited lots of anti-Latinx and anti-immigrant sentiment and hate crimes throughout the country). My point, rather, is that many of the arguments about generic statements end up missing these other points, because people interpret these statements in different ways and end up talking around one another.

The ability to make a strong statement while leaving yourself some wiggle room and ability to back out of that statement is one of the things that makes generic statements so powerful (and so dangerous). In fact, this characteristic is not unique to generic statements; as we will see in the "Benefits of indirect utterances" module near the end of class, this ability is one of the main reasons language has pragmatics at all.

Video summary

In-class activities

As we noted above, people are even more willing to accept a generic statement as true with little evidence if the statement is about a feature that is dangerous, distinctive, or "striking". In other words, people might normally accept "Morseths have silver fur" (this is another example from Cimpian et al., 2010) as true if actually 70% of morseths have silver fur; but if they are first told that morseth's silver fur is dangerous because it sheds air particles that can cause breathing difficulty, then they might accept "Morseths have silver fur" as true even if just 30% or even just 10% of morseths have silver fur! Lazaridou-Chatzigoga and colleagues (2019) further showed that even young children show this bias: children are more likely to make generalizations based on generic statements about "striking" features than on generic statements about "neutral" features.

Why do you think this happens? What communicative purpose do you think it serves? And how could this make generic statements extra dangerous (or, at least, extra prone to misinterpretation)?

⟵ Presupposition
Non-restrictive modification ⟶

by Stephen Politzer-Ahles. Last modified on 2021-11-12. CC-BY-4.0.