Inferential thought

TYPE OF PSYCHOLOGY: Cognition

An argument is a process that takes assertions as inputs and produces conclusions as outputs; to go beyond the information given and get from the inputs to the outputs is to draw inferences. Formal inferences include deduction and induction. Inferential thought of a formal and informal nature is essential to both scientific reasoning and reasoning in daily life.

Introduction

Psychologists are only beginning to understand how human thought processes operate, but there is no doubt that thinking is a critical skill. Reasoning is but one of many types of thought. Others include decision making and concept formation. Reasoning is unique in that it involves drawing inferences from current knowledge and beliefs. Reasoning has multiple components, including the production and evaluation of arguments, the drawing of inferences, and the generation and testing of hypotheses.

The process of inference involves the exploration of alternatives, using evidence. Evidence is information that helps determine the degree to which a possibility achieves a goal. Basically, using inference, each possibility for choice is made stronger or weaker, considering that goal. The process can be done well or poorly. Without the ability to make inferences, there would be no science, mathematics, or even laws.

Almost every statement a person says or writes leads the listener or reader to make inferences. A presupposition is knowledge on which one draws to understand a statement or assertion. Once the assertion is understood, an inference can be drawn. Certain types of inferences, known as logical inferences, must follow from what was said. Logical inferences are, in a sense, demanded by the assertions. For example, the statement “Jack’s heart problems forced his doctor to put him on a strict diet” logically implies that Jack was put on a diet.

Induction and Deduction

There are two basic forms of formal, logical inference: induction and deduction. An induction is a judgment that something is probably true on the basis of experience. It involves generalization—that is, reasoning from a few to all, or from the particular to the general. People infer that they should avoid all bees, having been stung by only one or two up to that time. The inductive inference allows one to go beyond the data at hand and draw a useful conclusion (all bees will sting people) that cannot be proven because it cannot be exhaustively tested. Induction can extend the content of the assertions at the cost of introducing uncertainty.

In contrast, deductive inference achieves absolute certainty, if performed correctly, at the cost of sacrificing innovation. It requires that two or more separate assertions be integrated to deduce a new assertion as a necessary consequence. Deductive inference deals with the validity, or form, of the arguments, providing methods and rules for restating given information so as to make what is implicit explicit. All valid deductive arguments reformulate knowledge already given in the assertions. They typically utilize key terms, such as quantifiers (such as “all,” “some,” “none”), connectives (such as “and,” “or,” “if-then”), and comparatives (such as “more,” “less”).

Inferring Sentence Meaning

An experiment published in 1972 by John D. Bransford, J. Richard Barclay, and Jeffrey J. Franks illustrated that people could not distinguish sentences that were actually presented from inferences they made in the process of comprehending those sentences. Subjects saw sentences such as “Three turtles rested on a floating log, and a fish swam beneath it.” Subjects were then given memory tests to see if they recognized logically implied sentences such as “Three turtles rested on a floating log, and a fish swam beneath them” that were new, so to speak, because they had not actually been seen by the subjects before the recognition test. A large number of the subjects claimed that they had seen the new sentences, which suggests that the logical inferences were formed and stored at the time when the original sentences were initially presented.

Not all inferences are demanded by formal logic, however; the majority of inferences are invited by the assertion, and they are known as pragmatic inferences. A pragmatic inference does not need to follow from an assertion, but rather is reasonable, considering world knowledge. For example, to say that “Albert and Pat were looking at wedding rings” in no way demands the inference that Albert and Pat are to be married; however, that inference is certainly reasonable, given what is known about the world. A large number of experiments have been reported that demonstrate that pragmatic inferences also are remembered as part of the original event.

As a further illustration, additional research by Bransford and other colleagues in 1973 presented subjects with sentences such as “John was trying to fix the birdhouse. He was pounding the nail when his father came out to watch him and to help him do the work.” The assertions imply, but do not logically demand, that John was using a hammer. Subjects later falsely recognized the sentence, “John was using a hammer to fix the birdhouse when his father came out to watch him and help him do the work.” Like the logical inference, the pragmatic inference is usually remembered as if it had actually been presented.

Shortcomings of Inference

Experimental investigations of thinking have revealed a wide range of shortcomings in human inference. Both deductive and inductive reasoning can go astray and produce incorrect conclusions, often either because one of the premises from which the conclusion was drawn is false or because the rules of deductive inference were violated. Many inferential judgments are based on imperfect information, and that means mistakes are unavoidable; however, the shortcomings are not simply errors. Instead, the ones that psychologists have identified involve the way in which information is used to draw the inference. For example, relevant information is sometimes ignored and sometimes relied on too heavily. In addition, multiple pieces of information are often not combined as they should be.

To understand human communication, it is necessary to recognize the prevalence and power of inferential processing. Much of what is communicated is actually left unsaid. Speakers instead rely on listeners to draw appropriate inferences. The ability to communicate without explicitly saying everything one is trying to convey enormously increases efficiency; however, as with other thought processes, increased efficiency comes at the cost of increased error. Everyone occasionally says things in such a way that a listener will infer information that may not be quite accurate. To determine whether a speaker is being dishonest, the speaker’s intentions need to be discovered, which is difficult if the actual assertion is accurate. As a result, it is easy to mislead—either when sufficient information to evaluate an assertion is intentionally withheld or when care in drawing inferences is not taken.

Implied Messages of Advertising

Real-world situations such as advertising copy and courtroom testimony provide interesting examples of the use of potentially misleading information. For example, the Federal Trade Commission was established to make decisions about what constitutes deceptive advertising, but deciding exactly what is deceptive is complex. The decision becomes especially difficult if a claim is not blatant but instead is implied. Consider the following commercial: “Aren’t you tired of sneezing and having a red nose throughout the winter? Aren’t you tired of always feeling under the weather? Get through the entire winter without colds. Take NuPills as directed.”

Notice that the commercial does not directly state that NuPills will get one through the entire winter without colds. The commercial only implies it. To test whether people can distinguish between asserted and implied claims, John Harris, in a study published in 1977, presented people with a series of twenty fictitious commercials, half of which asserted claims and half of which implied claims. The subjects in the experiment were told to rate the claims as true, false, or of indeterminate truth value, based on the presented information. Some people made their judgments immediately after hearing each commercial, and others made their judgments after hearing all the commercials. Half the people were given instructions that warned them to take care not to interpret implied claims as asserted ones.

The results were that the subjects responded “true” more often to assertions than to implications, and instructions did help to reduce the number of implications accepted as true. Overall error rates, however, were high. Even in the group that gave an immediate judgment after hearing each commercial, people mistakenly accepted about half the implied statements as asserted ones. Finally, when the judgments were delayed until all commercials were presented, people accepted about as many implied statements as true as they did direct statements, even when they had been specifically warned about implied statements.

Courtroom Testimony

In the context of how information can be misleading in courtroom testimony, Elizabeth F. Loftus published an article in 1975 that described how she showed subjects (witnesses) a film of a multiple-car accident. Immediately afterward, the witnesses completed a questionnaire that included questions such as “Did you see a broken headlight?” Half the witnesses, however, were given a question that was worded, “Did you see the broken headlight?” When the word “the” is used, the question encourages the subject/witness to assume that there was a broken headlight and seems to be asking whether they happened to see it. The word “a” does not presuppose the existence of a broken headlight. Questions with “the” more often led to reports that the witness had seen the broken headlight than questions with “a.” This was the case regardless of whether the object (a broken headlight) had actually appeared in the film. Thus, in a courtroom situation, attorneys can intentionally or inadvertently influence the memories of witnesses by using leading words that entail presuppositions in their questions, leading to inaccurate inferences.

Regardless of the source of the information, the elaborative nature of comprehension can be and is used to imply potentially inaccurate information. Yet through knowledge of influence, one can be in the position to protect oneself by directly questioning assertions and carefully analyzing one’s own inferences.

Evolution of Study

In ancient Greece, the philosopher Aristotle was the main creator of a formal inferential system. Historically, however, the scientific study of inference began fairly recently. Psychologists such as Robert S. Woodworth and S. B. Sells first began publishing articles in the 1930s on errors people made in the process of inferring conclusions. Woodworth and Sells were interested in how a reasoner’s personal attitudes toward the conclusion of a syllogistic argument could bias the ability to draw inferences. They, along with other psychologists during the next three decades or so, studied formal inference and mainly looked at logical arguments called categorical . As time passed, psychologists began to study other forms of deductive arguments and, later, inductive ones.

In 1962, Mary Henle encouraged psychologists to consider the difference between formal and practical reasoning. Henle attempted to clarify the heated controversy between the psychologists who thought formal logic was largely irrelevant to the thinking process, those who believed the mind contained a formal logic, and those who thought the mind contained other systems of logic that were more practical for day-to-day thinking. Henle pointed out that mathematical logic was never intended to be a direct description of how people think. For example, formal reasoning makes two demands not made in everyday reasoning. First, the reasoner must restrict the information to that contained in the premises. Second, the reasoner needs to discover the minimum commitments of the assertions as they are worded, which is not typical of ordinary comprehension. In ordinary comprehension, many inferences are invited that are unacceptable in formal deductive logic.

Linguists have also helped to promote research on inference. A long-standing question posed by linguists is how logic relates to actual conversation and argumentation. In conversation, most utterances have multiple functions. For example, the same utterance could be a description, a persuasion, an emotional expression, or even a warning. Numerous functions can arise because the speaker may have one of a variety of intentions in mind when making the utterance.

It has been argued by philosopher of language and reasoning Philip Johnson-Laird, in work he began publishing in 1978, that what subjects use to understand text is a mental model of the textual statements. He claimed that people construct representations of models when they read a text. Rather than relying on formal logic in the interpretation of the material, Johnson-Laird asserted, people manipulate the models they have formed. He noted that the psychology of reasoning should describe the degree of competence that people display when it comes to inference, but that the mental processes underlying them—inferential performance—also need to be investigated.

Given the increasing interest in the relationship of inference to linguistics, especially considering the applications that can be made in the area of , such as getting computers to understand speech and to translate from one human language to another, the psychology of inferential thought will undoubtedly continue to be an important area of research. Other important research areas include the role of the brain’s hippocampus in making inferences from integrated memories and improving inferential thinking skills among students using targeted strategies. Research indicates that students who learn to make inferences using contextual clues from a text and prior knowledge of the topic experience better comprehension of the lesson and develop critical thinking abilities. Teachers can foster this skill in their students by asking targeted, open-ended, and engaging questions during reading sessions.

Bibliography

Baron, Jonathan. Thinking and Deciding. 5th ed., Cambridge UP, 2023.

Evans, Jonathan St. B. T. Bias in Human Reasoning: Causes and Consequences. Erlbaum, 1994.

Evans, Jonathan St. B. T. The Psychology of Deductive Reasoning. Routledge, 1982.

Halpern, Diane F., and Dana Dunn. Thought and Knowledge: An Introduction to Critical Thinking. 6th ed., Routledge, 2023.

Hobson J.A., et al. “Dream Logic-the Inferential Reasoning Paradigm.” Dreaming, vol. 21, no. 1, 2011, pp. 1–15, doi.org/10.1037/a0022860. Accessed 29 Nov. 2024.

Laden, Anthony Simon. Reasoning: A Social Picture. Oxford UP, 2012.

Lane-Getaz, Sharon J. “Development of a Reliable Measure of Students’ Inferential Reasoning Ability.” Statistics Education Research Journal, vol. 12, no. 1, 2013, pp. 20–47, iase-web.org/documents/SERJ/SERJ12(1)‗LaneGetaz.pdf. Accessed 29 Nov. 2024.

Nickerson, Raymond S. Reflections on Reasoning. Erlbaum, 1986.

Rice, Marianne, et al. “Promoting Inference Generation: Using Questioning and Strategy Instruction to Support Upper Elementary Students.” The Reading Teacher, vol. 78, no. 2, 2024, pp. 121–30, doi.org/10.1002/trtr.2353. Accessed 29 Nov. 2024.

Sanford, David. If P Then Q: Conditionals and the Foundations of Reasoning. 2nd ed., Taylor and Francis, 2014.

Streumer, Bart. “Inferential and Non-Inferential Reasoning.” Philosophy and Phenomenological Research, vol. 74, no. 1, 2007, pp. 1–29, streumer.net/InferentialReasoning.pdf. Accessed 29 Nov. 2024.

Wilson, Deirdre, and Dan Sperber. Meaning and Relevance. Cambridge UP, 2012.