Problems with word definitions



Back to main page

Back to home page

Printer friendly

Problems With Word Definitions

Word definitions became a national issue a few years ago when Bill Clinton tried to use some tricky definitions to avoid revealing his relationship with Monica Lewinski. He claimed he "did not have sexual relations with that woman" but it turns out he clearly did have a sexual relationship with that woman. He tried to play games with the definition of "sexual relations" (since he didn't actually have intercourse with her) but it just got him into deeper trouble. By the same token, responsible thinkers should be careful about using definitions improperly.

What determines the real meaning of a word?

People argue a lot about the meanings of words. How can these arguments be resolved? What is the ultimate authority when it comes to word meaning? Normally we think of dictionaries. But there are a lot of different dictionaries with slightly differing definitions, and there is no reason to say that the person who wrote a particular dictionary definition is necessarily correct. So how do dictionary writers determine what the meaning of a word is? The answer is that they determine it from common usage - people using the words in the course of what they say and what they write. Usage changes over time and is likely to vary somewhat from place to place and person to person. So it turns out that the ultimate authority for a word might not be some linguistics professor who writes for the Oxford English Dictionary, but some punk rocker who uses the word in a certain way that catches on with the public so it becomes common usage. The point is that it usually doesn't make sense to say a word really means this or that particular thing. Fine points of word meanings are almost always going to be slippery.

This is even true for something as simple an obvious as a "chair". An obvious attribute of a chair is that it is something to sit on, and it usually has a back to lean against. However a sofa or a loveseat aren't called chairs, presumably because they are wide enough for more than one person. Stools and hassocks are not called chairs because they don't have backs. But some bars stools have backs, but are still called stools instead of chairs. That's probably because they were modifications of stools that didn't have backs rather than modifications of chairs that were shorter, so the name "stool" continued to be used. The "chair" lift at a ski resort is likely to have something more similar to a bench than a chair. A beanbag chair is not at all like a stereotypical chair. When we look at the actual usage of the word "chair", we find that there is no simple underlying meaning that is totally consistent, but we all have heard the word used often enough that we are familiar with most of the oddities of when it does and does not apply.

The fact that words don't have fundamental underlying meanings creates problems for language translation. It may be that in some other language, the word with the meaning closest to chair includes sofas, and it another language it might exclude sofas but include chairs without backs that we would call stools. Translators will try to pick the closest word that fits the context, but when we read the translation, we may still misunderstand what the original author intended.

There are some contexts where word meanings actually are relatively precise. Science is one example. Scientists go to great pains to define terms to avoid any ambiguity. Scientific quantities are often defined in terms of objective measurements that minimize the chance for differences in interpretation by the people doing the measuring. This precision in defining terms is probably one of the key reasons science has been so successful over the years.

Another area where terms are carefully defined is law. Words used in laws and contracts have to be carefully defined because the system will fail if people don't know exactly what is required.

Definition issues that mask substantive issues

For most purposes, though, we have to be careful about getting hung up on definitions. When people think that a definition is something worth arguing about, we should be alert for a problem.

At one time I was curious about the definition of the word "disease". People seemed to take great pains to point out that "alcoholism" was a "disease". I wondered, what difference does it make? The controversy didn't seem to be about the nature of alcoholism so much as whether alcoholism was included in the definition of "disease". A person could easily define the word "disease" to include or exclude it - alcoholism was still the same thing. Eventually it dawned on me that the debate was really about how we should treat alcoholics: in the past we might have ridiculed or jailed them, but if we can persuade people to think of alcoholism as a disease, putting them in hospitals or treatment programs seems more appropriate. There was actually an issue of substance at the root of the question of the definition of disease. For responsible thinking purposes, it seems much better to address the issue of substance directly. The real issues should be the effectiveness, and the cost, and the humane considerations of each approach. We don't learn anything from debating the semantics of the word "disease".

Another example is the definition of the word religion. One issue that will always result in a heated debate among people who consider themselves "secular humanists" is the question of whether secular humanism is a religion. There isn't too much dispute about the actual nature of secular humanism - it tries to promote the betterment of humanity without belief in God or the supernatural. Is that a religion? Must religions always involve a belief in the supernatural? Most of them do but the word religion could be defined in such a way as to either include or exclude secular humanism. But it seems to me that the debate isn't really about what the word "religion" means. It is actually about the public relations consequences of calling it a religion. Some secular humanists want to dissociate themselves from the features they dislike in other religions, while others want to benefit from some of the positive associations the public typically has with religion. There really is an issue of substance here that tends to get clouded because the debate seems to be about semantics.


In addition to cases like the above where imprecise meanings can cloud the real issue, there are many situations where an argument is invalid because the meaning of a word changes within the argument. This is sometimes called equivocation.

Some people might say that when a man shows a sexual interest in a woman he is treating her as a "sex object" and therefore, since she is a "object" to him, he doesn't think of her as a person. This isn't a valid argument because the word "object" in "sex object" doesn't deny humanity as it would if we were talking about an inanimate "object". The term "sex object" was used in psychology to refer to whoever a person was sexually attracted to, which was fine if it was their spouse, but bad if it was their mother or father. Many people might not notice the flaw in this argument because it is easy enough to think of cases where men do show disrespect towards women they pursue for sex.

Another argument involving equivocation might have go like this: A dictator has weapons of mass destruction. Weapons of mass destruction can destroy entire cities. Therefore the dictator has weapons that can destroy entire cities. The problem is that "weapons of mass destruction" effectively has three meanings corresponding to the type of weapons: nuclear, biological, and chemical. Of these, only nuclear is has a realistic capability to destroy a city. If the only "weapons of mass destruction" the dictator has is chemical (which, in the case of Iraq, seemed the most likely), then he really doesn't have the power to destroy cities.

Here is another argument that is faulty because of equivocation: Thomas Jefferson was a Christian. Christians believe Christ rose from the dead. Therefore Thomas Jefferson believed Christ rose from the dead. As it turns out, Jefferson did not believe this. He was not a Christian in the ordinary sense. He regarded Jesus as a great philosopher and considered himself a Christian because he believed in most of the teachings of Christ; but he did not actually believe in the miracles attributed to Christ or the resurrection.

Distorting definitions to promote an agenda

Sometimes people with a political or philosophical agenda like to define a word in a way different from common usage in order to advance a cause they favor. They justify this by using the root meaning of the word or some historical meaning. Here are a couple of examples:

I have heard it argued that something that is "supernatural" must be non-natural, and since any phenomenon that actually exists is natural, nothing supernatural can possibly be true. Therefore, it might be reasoned, ghosts cannot exist. In reality, when people think of the supernatural, they think of certain things that have traditionally been called supernatural, such as ghosts, ESP, and magic. They may or may not think these things cannot actually occur. There would not be much point in having the word supernatural if it inherently meant things that can't occur. We might as well just use the word "impossible". But even if we did use the word that way, it wouldn't imply ghosts couldn't exist - it just would imply that ghosts, if they existed, were "natural".

Another claim of this sort is the claim that everyone is "selfish" because, even if you spend a lot of your time helping other people, you only do it because you like to, so you are really doing it for selfish reasons. Once again, this is not how people ordinarily use the word - people who helps others because they like to are normally thought of as unselfish. As in the "supernatural" example, the new definition makes the word essentially meaningless. After establishing the idea that everyone is selfish, the argument may go on to imply that this has implications for economic policies. However it can have no implications, because the selfishness involved is automatic by definition rather than being an observed characteristic of human psychology, and so it really tells us nothing about how people will behave given actual economic conditions.

Formal definitions

Formal definitions are definitions made artificially for a specific situation or within a specific context. For example, a person analyzing poetry could define a "double" as a pair of lines beginning with the same letter of the alphabet. Within the context of the person's analysis, that is what "double" should be taken to mean, even though that is not an ordinary meaning from usage. This sort of definition can be useful for a variety of situations and so is a legitimate thing to do. However this also can be abused.

I have heard of the word "poison" being given a formal definition of "any substance that does not naturally occur in the body". Therefore any drug such as aspirin which people may take is categorized as "poison". We have to remember that, while it is valid to redefine a word this way, once it is redefined the old definition, in this case something that could kill you if you eat it, is no longer relevant. This definition was promoted by people who think we should only ingest "natural" substances, and so they quite clearly wanted to create the link between unnatural substances and deadly ones, but redefining a word is not a valid way to show such a link exists.

If we want to be responsible thinkers, we should put a low priority on debating purely about definitions, since the ultimate source of word meanings, common usage, isn't static or consistent. We should look carefully at arguments about definitions to see if an issue of substance might really be at the root of the debate. In addition, when evaluating arguments (particularly our own), we should be careful that definitions don't change from one part of the argument to another and that definitions aren't contrived to suit the purpose of supporting a preconceived position.