Generalizations can be a valid method of argument. Inductive reasoning, in particular, is based on the ability to generalize from repeated experiences or observations. The soundness of an inductive generalization can usually be determined by asking the following questions:

  1. Do we have a sufficient number of instances to draw a conclusion?
  2. Is the breadth of the conclusion drawn supported by the evidence?
  3. Are the terms of the conclusion consistent with the terms of the evidence?
Fallacies result if any of these questions can be answered in the negative.


A hasty generalization is one in which there is an insufficient number of instances on which to base the generalization. Consider the following examples:

  1. Jana has been to San Diego several times, and the sky was always blue and the temperature ideal. The weather must be perfect in San Diego all the time.
  2. Tina bought a used camera while she was up in Portland, and got a great deal. Portland must be a good place to buy used cameras.
  3. I read where there have been no reported cases of HIV infection in Liberty Lake. The people of Liberty Lake must be free of the HIV virus.
In the first two examples, generalizations were made on the basis of little evidence--several days in San Diego, one camera purchased in Portland. These clearly provide an insufficient basis for the conclusions they are used to support, and are therefore examples of hasty generalizations.

The third example is a little different. There, a generalization is made on the basis of no evidence at all. The lack of evidence to the contrary should never be used as sufficient grounds for any generalization. For example, the absence of a suspect's fingerprints on the murder weapon is not sufficient in itself to prove his innocence, nor is the lack of any evidence of life in soil samples taken so far on Mars sufficient in itself to prove that no life exists there. This is a special case of hasty generalization, usually known by its Latin name, argumentum a silentio, or argument from silence, because instead evidence to support the argument, all we hear is silence.

The problem in each of these cases should be obvious: without more data, we have no way of knowing if the evidence presented is representative or not. Maybe Jana happened each time to visit San Diego during unusually good weather, maybe Tina was really lucky to get a good deal on the camera, maybe people are reluctant to reveal that they are HIV-positive in Liberty Lake. Without sufficient support for the generalization, these are just anecdotes.


A sweeping generalization is one in which there seems to be sufficient evidence offered to draw a conclusion, but the conclusion drawn far exceeds what the evidence supports. Consider these examples:

  1. The profit margin on HP's printer line has been a steady 25% for two years. We can assume, then, that the profits company-wide have also been 25%.
  2. The poll from Orange County shows the governor winning in a landslide. I guess he will also win across the state just as easily.
In each example, the conclusion drawn far exceeds what the evidence would support. For all we know, the printer line is part of HP's profitable personal computer division, and we might be able to extend the findings to similar products in HP's line, but not to the full line itself without a great deal more information. In the second example, we could certainly conclude that the governor will win in Orange County, and perhaps we might be willing to conclude that the governor should be favored in similar counties, though the nature of the similarity may not, at first, be very apparent--if geographical, demographic, political, and economic, and so on. But assuming that the entire state is somehow similar to Orange County, which is an assumption that you would have to accept to make this argument, is stretching the evidence of similarities well beyond the limit.


The third question about a generalization asks about consistent terms. Consider the following examples:

  1. I used only delicious ingredients, so this sauce must be delicious.
  2. The 49ers are the best team, so they must have the best players.
The problem in both is that non-equivalent terms have been substituted: the parts (ingredients) for the whole (sauce) in the first example, and the whole (team) for its parts (players) in the second. Notice how this fallacy usually involves the replacement of a plural noun, such as "ingredients" and "players," with a singular, collective noun, such as "sauce" and "team." And, generally speaking, the whole (the collective noun) is often more or less than the sum of its parts (the plural noun). Substituting the whole for its parts, the sauce for its ingredients, is sometimes called the fallacy of composition. Substituting the parts for the whole, or the players for the team, is sometimes called the fallacy of division.

Exercises

1. Which of the following is not a fallacious generalization?

That restaurant has been closed every time I've tried to eat there. It must never be open!

That restaurant has been busy every time I've tried to eat there. It must have delicious food.

That restaurant has been closed by the health department. The inspectors must have found code violations.

That restaurant has been given excellent reviews by every newspaper in town. It must have an excellent staff.

2. Which of the following is not a fallacious generalization?

All generalizations are fallacious, including this one.

This is a research library, so all its librarians must be involved in research.

Of course I believe in Martians, since there's no evidence that they don't exist.

All trees are plants, including redwoods.

2. Which of the following is not a fallacious generalization? You answered:

All generalizations are fallacious, including this one.

Well, this one is, because not all generalizations are fallacious. This would be a sweeping generalization, since what is at issue is not the number of examples on which it is based, but its scope. There are, certainly, plenty of examples of fallacious generalizations. But to conclude on that basis that all generalizations are fallacious ignores plenty of other examples of reasonable generalizations.

2. Which of the following is not a fallacious generalization? You answered:

This is a research library, so all its librarians must be involved in research.

An example of substituting the parts for the whole. That a company is productive does not mean that each of its workers is productive; that a forest is old growth does not mean each tree in it is old growth. So this generalization about librarians in a research library is fallacious.

2. Which of the following is not a fallacious generalization? You answered:

Of course I believe in Martians, since there's no evidence that they don't exist.

This is an argument from silence, because it draws a conclusion ("since") on the basis of a lack of information ("no evidence"). This particular example might also be considered a fallacy of shifting the burden of proof, since the subject here is one where the burden of proof would naturally fall on the side that believes in Martians.

2. Which of the following is not a fallacious generalization? You answered:

All trees are plants, including redwoods.

Correct!

Many generalizations, such as this one, are not fallacious.

1. Which of the following is not a fallacious generalization? You answered

That restaurant has been closed every time I've tried to eat there. It must never be open!

This argument contains the generalization that, because the place was closed every time I was there, it is always closed. This is probably a hasty generalization (unless I've been there all day every day for a substantial period of time). Notice that this generalization may be true--but we don't have sufficient support to reach that conclusion.

1. Which of the following is not a fallacious generalization? You answered

That restaurant has been busy every time I've tried to eat there. It must have delicious food.

This argument contains two fallacious generalizations: (1) that all busy restaurants have delicious food (it could be the prices, or the location); and (2) that because it's busy every time I've tried to eat there, it's always busy. The first is a sweeping generalization (there's some connection between a busy restaurant and its food, but not enough to make that blanket statement). The second is a hasty generalization (unless I've been there all day every day for a substantial period of time). Notice that both of these generalizations may be true in this case--but we don't have sufficient support to reach those conclusions.

1. Which of the following is not a fallacious generalization? You answered

That restaurant has been closed by the health department. The inspectors must have found code violations.

Correct!

This argument is based on the reasonable generalization that when a restaurant is closed by the health department, violations of the health code must have been found. Not all generalizations are fallacies.

1. Which of the following is not a fallacious generalization? You answered

That restaurant has been given excellent reviews by every newspaper in town. It must have an excellent staff.

There are two generalizations here: (1) restaurants that receive many excellent reviews are excellent restaurants, and (2) excellent restaurants have excellent staffs. The first generalization seems to be reasonable enough--at least as far as any reviews can determine the quality of the thing reviewed--but the second substitutes the parts (the staff) for the whole (the restaurant), and so becomes a fallacy of substitution. After all, the reviewers may have overlooked bad service because of superb food or an interesting decor.

Congratulations!
You have finished the exercises for the "Fallacious Generalizations" section.
Click here to return to the menu.