Extracting Arguments

Extracting Arguments

It should come as no surprise that we will be reading a great deal of philosophy this semester. Reading philosophy is not like reading the Sunday comics, although some wish it were. When philosophers write, they try to convince the reader that their point is true (or, just as often, that some other philosopher's point is false.) In order to convince their readers, philosophers give reasons. But reasons can often be very complicated. Worse, the connection between their reasons and the philosopher's point may not be altogether clear.

I take it as one of my duties in this course to try as best I can to explain what a particular author is saying. With all due credit to Fred Feldman, what this means is that I first extract an author's arguments, I then explain the author's arguments, and I conclude by evaluating the author's arguments. Together these steps make up a thorough critical analysis. Since you need to learn to do this for yourself, let's break the process down a bit further.

First, extract the author's arguments:

Any author worth reading will have a point or thesis to their writing. The first thing to do is figure out the thesis (or theses, as the case may be.) Having read the article, ask yourself "what point or points is the author trying to get me to accept?" Good writers will alert you to their point/thesis or points/theses by using special words or phrases as signposts. We call these words or phrases 'conclusion markers'. There are many conclusion markers. Examples include 'therefore', 'in conclusion', 'it follows that', 'hence', and 'accordingly', among others. Since your goal is to write as clearly as possible in your cases, you should learn to use these words yourself.

Once you've expressed each of the author's theses as a single statement, your goal is to uncover the reasons the author gives for thinking that his or her theses are true. This is often very hard to do. Good writers will take pains to signpost their reasons with so-called premise markers. 'Since', 'because', 'assuming that', 'granting that', and 'given that' are some examples of premise markers. But even good writers will often not explicitly state all of their reasons. Your task is to uncover the reasons they explicitly give, and then try to figure out what reasons they must be assuming implicitly. This is hard to do, but it is not groping in the dark. In extracting the author's argument or arguments, we assume that the author was smart enough to not give us any invalid argument. That is to say, we grant the author the presumption that his or her arguments, if deductive, can be expressed in valid form. But putting an argument in valid form almost invariably reveals 'holes' in the argument or premises (reasons) which the author has not seen fit to supply. We then supply them.

By the time we finish extracting an author's arguments, we should have, if we did it correctly, an argument or series of arguments in valid form.

Second, explain the author's arguments:

The point of this step is to be sure that we understand all of the terms the author is using so that we understand what the arguments really state. In particular, we want to be sure that we understand what each of the premises is saying. A good way to approach this, and a technique you'll often see me using in class, is to try to figure out what conditions or facts would have to obtain for the premises to be true. In other words, we have a good shot at understanding the premises if we can clearly state the grounds for the truth of the premises. This is not to say that the premises are true, of course. The grounds or facts simply may not obtain.

Third, evaluate the author's arguments:

The idea behind this step is simple, but it is probably the hardest step to pull off correctly. Recall that we now have before us a valid argument or a series of valid arguments. Remember, too, that an argument is valid if it is in valid form and to be in valid form it must be the case that the premises cannot all be true and the conclusion false at the same time. So if the premises are all true, then the conclusion must be true. We know that the author's (deductive) arguments are valid, because we've given him or her the benefit of the doubt and forced the (deductive) arguments (sometimes this is a stretch, of course) to be valid. What we want to know, then, is whether or not the author's arguments are sound. Really that's all this step amounts to. We have these valid arguments before us, so now what is left is to try to figure out whether or not the premises are true. If the premises are all true, then the argument is sound and its conclusion is inescapably true. (I say 'inescapably', since it is sometimes the case as we will certainly see this semester that an argument is found to be, to the best of our knowledge, sound, but we don't like the conclusion at all. Still, if the argument is in fact sound, then we are driven to the conclusion whether we like it or not.)

So, this step just amounts to establishing the soundness of the arguments. Our approach is critical: we begin with the presumption that at least one of the premises in each argument is false. The problem is to figure out which one. Sometimes we have to supply counterarguments--i.e., further arguments which we are confident are sound which conclude with the negation of the offending (false) premise. This makes sense if you keep in mind that the negation of a false sentence is true. Other times we have to supply counterexamples--i.e., facts which obtain but could not if the premise was true.

If the author's arguments make it by our best efforts to show that one or other of their premises are false, then we grudgingly admit that the author's arguments are sound, at least to the best of our knowledge.

If the author's arguments fail to make it past our best efforts and we succeed in showing that each argument has a false premise, then we conclude that the author has failed to show that their thesis or theses should be accepted. What we do not conclude is that the author's thesis or theses is false. Even if all the author's arguments are shown to be unsound, it is still possible for the author's thesis or theses to be true. This makes sense only if you recall that it is possible for a valid argument to have a false premise and a true conclusion. Oftentimes, then, you will take the additional step of giving counterarguments to show that the author's conclusions are not true. But in the absence of counterarguments, the best we can do is shrug our shoulders and admit that we don't know whether or not the author's conclusions are true. At the very least, we know in this case that the author has failed to give us any good reason for thinking that his or her conclusions are true.

Like it or not, that's the process. Believe it or not, that's what it takes to read an article carefully. Needless to say, most people aren't prepared to read anything very carefully. Most of the time that's ok: the issues aren't often all that important. But when we talk about things like abortion and euthanasia, we are talking about very important issues--issues people are sometimes willing to kill each other over. We owe it to these issues to do everything we can to get at the truth. If that's too ambitious, at least we can discard clearly false positions (much as we discarded clearly false ethical theories.)

Clearly, the key to reading philosophy carefully is the extraction step. If we have not successfully extracted an argument, then it is impossible for us to say anything intelligent about what the author has to say. Extracting arguments is something of an art. Since arguments occurring in written or verbal passages are often obscure, and are rarely as neatly arranged as our examples have been, certain skills are needed in order to interpret what is being argued for and from what assumptions. The following, adapted from Copi and Cohen's "Introduction to Logic" (New York: Macmillan, 1994), 9th ed., are a series of points to bear in mind when attempting to extract arguments:


The conclusion is always the result of an inference drawn from the premises; therefore, the conclusion in an argument is always what is argued for, while the premises are taken for granted or assumed to be true. We say the conclusion is inferred from the premises. The conclusion is the single sentence in the argument requiring justification. Looking at a passage, you should always first find the sentence whose truth is not taken for granted but whose truth is the whole point of the passage.


Premises and conclusions are always relative to a single argument. What is taken as a premise in one argument may be the conclusion of another argument. Arguments in long passages may therefore be linked one with another.


Typically, the statements in an argument will be in the form of declarative sentences, but not always. Sometimes questions, particularly rhetorical questions, and commands will occur as premises. This will happen only when such sentences can be thought of as making a claim even though they may not be in the form of assertions. Any sentence, regardless of its mode, that can be interpreted as making a claim is said to express a proposition. The claim made is said to be the proposition expressed. This is perhaps the simplest way to think about what a proposition is. One can also define a proposition as what any two synonymous sentences, whether they be in the same language or in different languages, express. Any sentence that may be said to express a proposition may be a premise or a conclusion in an argument.


In the logical analysis of arguments, we prefer to restate the premises and conclusion in such a fashion that the proposition expressed by each sentence is explicit. But, the meaning of a sentence is quite often dependent on its context. To understand what proposition is being expressed, you need the information provided by the context. Pronouns occurring in a sentence sometimes have their antecedents outside the sentence, and verbs have tense. The combined effect of these factors is such that the time or place in which the sentence occurs determines the proposition expressed. Take, for instance, the sentence, "The current President of the United States is a Republican." This sentence would be false when written in 1978, but it was true in 1990. Therefore, the context (in this case, the time at which the sentence is uttered or written) provides the exact proposition expressed. As another example, consider the sentence, "Hitler died in Berlin." This sentence, had it been uttered in 1932, would be false. If it had been uttered any time after 1945, it would be true. In the logical analysis of arguments, we prefer sentences that are as little dependent on context as possible. Such sentences express their propositions more explicitly than sentences that are more context dependent. So, for example, the sentences above might be restated in the following manner. "The President of the United States in March of 1990 is a Republican." "Hitler dies in Berlin on April 29, 1945." The ideal of logical analysis is always to work with sentences that explicitly (as much as possible) express their propositions. Hence, when we reconstruct an argument by restating its premises and/or its conclusion, we shall favor declarative sentences in simple present tense that contain no pronouns and time indexicals such as "current," "present," "past," "recent," "future," etc.


So far we have given arguments whose conclusions occur right after their premises. In an argumentative passage, the conclusion may actually occur anywhere in the passage. In some arguments, their conclusions are stated first. With others, their conclusion is stated in the middle, and for still others, it is stated at the end. The placement of the conclusion does not affect the argument's validity or soundness.

The premises and conclusion may be conjoined in a single sentence. The single sentence may be a compound sentence but may also be a complex sentence. Or the sentence may contain a noun phrase, which itself expresses a proposition even though it does not contain a subject and a verb. The point here is that the premises and conclusion may occur as parts of sentences that are not themselves sentences. So, for example, consider this argument:

If numbers were symbols, then there would be just as many numbers as there are symbols. However, there are only a finite number of symbols and an infinite number of numbers. Hence, numbers are not mere symbols.

You will notice that the second sentence is not a compound sentence even though two propositions are expressed in the sentence. It is saying that there are only a finite number of symbols and an infinite number of numbers. What we want to do in analyzing an argument is to restate it in such a fashion that we know exactly what the premises and conclusion are and that each is stated in a single declarative sentence. Thus, this argument may be rewritten in the following form.

  1. If numbers are symbols, then there are just as many numbers as there are symbols.
  2. There are only a finite number of symbols.
  3. There are an infinite number of numbers.
  4. Therefore, there are only a finite number of symbols and there are an infinite number of numbers.
  5. If there are only a finite number of symbols and an infinite number of numbers, then it is not the case that there are just as many numbers as there are symbols.
  6. Therefore, it is not the case that there are just as many numbers as there are symbols.
  7. Therefore, numbers are not symbols.


In an argumentative passage, certain words or expressions are particularly helpful in identifying which statements serve as premises and which statement is the conclusion. Perhaps the most frequently used premise indicators are the words "since" and "because," while the most common conclusion indicators are the words "hence" and "therefore." The following argument provides an example.

If God were both wholly good and omnipotent, then he would intervene in human affairs to prevent needless suffering. Since God is omnipotent, we can conclude that he is not wholly good because he never intervenes in world affairs to prevent unnecessary human suffering.

Notice that this argument contains the premise indicator "since" and the conclusion indicator "we can conclude that." Writing this argument in standard form, we have the following:

  1. If God is both wholly good and omnipotent, then he intervenes in human affairs to prevent needless suffering.
  2. God is omnipotent.
  3. God never intervenes in world affairs to prevent unnecessary human suffering.
  4. Therefore, it is not the case that God is both wholly good and omnipotent.
  5. If it is not the case that God is both wholly good and omnipotent, then either God is not wholly good or God is not omnipotent.
  6. Therefore, either God is not wholly good or God is not omnipotent.
  7. Therefore, God is not wholly good.

Don't, however, rely only on these indicators; they may not always be present an argument, and these same words and expressions may not always be used to indicate a premise or a conclusion. Keep in mind, particularly in the absence of these indicators, that the conclusion will always be what is argued for in an argument while the premises are what justify the conclusion and are assumed to be true. The ability to recognize this structure of an argument is actually more important than being able to pick out the premise and conclusion indicators, although the latter are indeed useful.


An argumentative passage may contain information that is not part of the argument proper. A good example is given by the following passage:

Bertrand Russell made the following case against naive realism. Naive realism, he claimed, leads to physics, and if physics is true then naive realism is false. Hence, naive realism is false.

The first sentence is not really needed, for its content does not help establish the conclusion. It only tells you that what follows is an argument. This argument, properly speaking, begins with the second sentence. We can rewrite the argument in our standard format as follows:

  1. If naïve realism is true, then physics is true.
  2. If physics is true, then naive realism is not true.
  3. Naïve realism is true.
  4. Therefore, physics is true.
  5. Therefore, naïve realism is not true.
  6. Therefore, naïve realism is not true.

Please note that this is an example of a Reductio ad Absurdum argument.


Occasionally a premise, or perhaps even the conclusion itself, may be left out of the argument. Even though they are not explicitly stated, the missing premises or conclusion are still essential components of the argument. The context in which the argumentative passage occurs may allow the author either to forego an explicit statement of a premise or to merely suggest a conclusion. Generally, there are two distinct circumstances in which this occurs. The author can make an appeal to common knowledge and leave the premise or conclusion implicit in the argument m virtue of the fact that any reasonable person would acknowledge it. In other words, the premise or conclusion is so obviously true that the author does not bother to state it. The other case involves a premise that is itself a conclusion of an earlier argument and is then not restated in a later argument. The following passage from Benjamin Franklin's Poor Richard's Almanac, 1746, illustrates this fact.

Dost thou love life? Then do not squander Time; for that's the stuff Life is made of.

You will notice that the first sentence is a question. Franklin is assuming, of course, that everyone loves life. This question is merely rhetorical. He is also assuming that no one squanders what they love. The complete argument in this passage can be rewritten as follows.

  1. Everyone loves life.
  2. Life is time.
  3. Therefore, everyone loves time.
  4. No one should squander what they love.
  5. Therefore, No one should squander time.