Explanation, in a narrow sense, is based on logic, which is study of the methods, and principles used distinguishing the valid from invalid reasoning. Besides sense-experience, every science is based also on the principles of logic or reason. Science involves the rules of reasoning or use of arguments. Arguments are sequences of sentences in which some sentences (premises or antecedents) are claimed to be the justification or support for another sentence (conclusion or consequences).
This justification or support offered on the basis of connection, relationship, association, property, common variable or attribute found between things and activities mentioned in the argument. Their source, ultimately, is sense-experience or empirical observation made by man in his social environment.
Logic is interrelated set of arguments. Arguments relate to validity and truth. Arguments can be (i) deductive, and, (ii) inductive. It should be pointed out that validity and truth are two different things. Validity means that the premises of an argument are related to the conclusion in such a way that the conclusions must be true if the premises are true. Validity is the property or attribute of arguments.
It says nothing about the truth or falsity of individual statements. It the premises of a valid argument are true, then the conclusion must automatically be true. Truth is a property of individual statements which can be either premises or conclusions. It is meaningless to call individual statements ‘valid’ and also to call arguments ‘true’. Arguments can be valid or invalid. Similarly, only individual or single statements can be true or false.
Deductive argument is one whose premises provide conclusive evidence for the conclusion. If all the premises are true, the conclusion must also be true. Deductive arguments proceed from general premises to more specific conclusions. The truth of the conclusion is implicit in the premises. It makes the information contained in the premises explicit. Inductive argument derives a generalised conclusion on the basis of particulars, often empirically derived premises or observations.
The premises of an inductive argument make the conclusion probable not certain. The inductive approach relies on scientific discovery of facts. One characteristic of inductive argument is that it establishes a conclusion whose content goes beyond its premises.
From the observation of a sample, an inference is made about a whole population. This is called the ‘inductive leap’, jumping from the premises which relates to an observed sample, to the conclusions which concerns the entire population. The greater the number of representative units in the premises observed in the sample, the smaller is the inductive leap.
The smaller the number of representative units observed in the sample, the greater is the inductive leap. The premises of an inductive argument do not establish the conclusion conclusively. The premises of a valid argument, here, may be true, but the conclusion can still be false. Its premises only support or lend weight to the conclusion, as they do not make the latter certain.
On the basis of two types of arguments or reasoning, explanations fall into two categories:
(a) Deductive explanation, and,
(b) Inductive explanation.
Explanations are, basically, answers to ‘why’ questions. In fact, there are many other styles of explanations: genetic, rational, dispositional, and functional. Genetic explanation gives the history of an event by presenting the event as the final stage of a development of a developmental sequence. The event may be described as happening through the hands of many specific individuals or agents directly or indirectly connected with it.
In fact, historical description of events cannot explain why that event took place. However, the events forming the historical sequence are selected on the basis of what the scholar assumes the cause of the final event. These assumptions form his tacit generalisations. The description of preceding events may be regarded as statement of particular facts or initial conditions.
Rational explanations take into account the agent’s calculation of means available in a situation for the attainment of his objectives. His actions are explained by listing the reasons that motivate him to act as in decision-making and communication theories. A decision-maker’s behaviour is explained by calculating the utilities or pay off under different conditions and circumstances. Rational explanations like genetic explanations, contain tacit generalisations which can be treated as covering laws.
But they suffer from two shortcomings: their generalisations actually are hypotheses or conjectures, and are not confirmed regularities, and, they often assume that the required initial conditions obtain without independent evidence. Dispositional explanations take into account person’s dispositions, tendencies or attitudes making them to behave in some particular manner. It can be assumed that man happens to be authoritarian, civic-minded, peaceful, violent or non-violent owing to them.
Frequently such orientations are used to explain why peoples and groups behave in a particular manner. All such assumed dispositions can be reconstructed as covering-laws. Only research or empirical observations can confirm whether man actually carries those dispositions under the given circumstances. Functional explanations take into their consideration the recurrent activities, behaviour-patterns or roles of persons in maintaining a system or in realising a particular goal.
Functional analysis examines a system that requires certain conditions or requisites to remain in adequate working order and particular persons or groups behave to maintain those conditions. It seems that the behaviour of these persons or groups satisfies a need or a functional requirement of that system.
Here the functional explanation takes the form of extended affirming of the consequent arguments, affirming or not affirming the antecedent conditional arguments. The antecedent conditional is placed in the form a covering law. It leaves little scope for ‘functional equivalents’, and does not properly define the terms, and assumes antecedents without empirical examination of facts.
In essence, they all are of similar nature, and can be subsumed under two aforesaid categories. A scientific explanation includes a set of laws or generalisations in the premises. As they cover the situation or event to be explained, these are called ‘covering laws’. Explanations that use covering laws are called ‘covering law explanations’ or nomological explanations.
They appear in two forms: deductive and inductive.
Deductive explanation is very powerful because it makes use of a valid form of deductive argumentation where the explanandum must be true if the explanans are true. Both individual and general events can be explained with certainty as the covering laws and the initial or given conditions are true. But such explanations have limited applicability to political and other social sciences.
In Political Science we do not have universal covering laws. Understanding of deductive explanation helps us to understand even inductive explanation because of their similarity in a number of respects. However, political scientists cannot give up the search for universal generalisations or covering laws.
In some areas deductive explanations have been evolved and achieved. Analysis of deductive explanations is necessary because of their employment by ideologues and spokesmen of various causes, interests, and power-patterns. Analysts should remain ready to challenge validity as well as truth of deductive arguments and explanations.
Inductive explanations, like deductive explanations, also have explanandum and explanans. Political Science abounds in inductive explanations which contain statistical generalisations in their explanans. Explanandum, ther
efore, cannot be deduced from the explanans with certainty. The explanandum is generally probable. It is implied by the explanans.
The explanans confer support or evidence for explanandum but does not make the latter certain. The explanans can be true and the explanandum can still be false in inductive explanation. Inductive explanations explain either the probability of individual events or statistical generalisations.
Major difference between deductive and inductive explanations is that one negative event can falsify a deductive explanation, but one negative event cannot falsify an inductive generalisation. Deductive explanations usually employ universal generalisations. Only some negative mass events or a high frequency of negative events can falsify inductive explanation.
It may be pointed out that explanations both in social and natural sciences are rarely presented in formal or standard logical form. Generalisations and initial conditions are not explicitly specified. In most cases, they are either assumed or treated as implied or understood. Demand of readability too compels scholars to do away with formal forms.
Many explanations presented in paragraph style conceal unspecified and untested generalisations or initial conditions. They can be reconstructed in proper manner and presented systematically exposing their logical errors and limitations. On the basis of degrees of explanatory completeness, there can be (a) complete, (b) elliptic, (c) partial, and (d) sketchy explanations.
In complete explanation all relevant generalisations and initial conditions are completely known and completely specified. But in Political Science, such explanations remain as ‘ideals’. In elliptic explanations, all the generalisations and initial conditions are known but all of them are not specified. The scholar assumes that the reader already knows them all and does not mention. But this omission is harmless and can be recouped easily. In the partial explanations all the relevant generalisations and initial conditions are not known. Only the known ones are specified.
As the relevant generalisations and facts or conditions are not given, the claimed explanandum cannot be deduced from the explanans. In other words, the explanandum appears to claim more than what appears in the explanans. The sketchy explanations or Explanation Sketches do not contain all relevant generalisations and conditions. Arguments are advanced on the basis of some hypotheses or stray incidents or conditions. Explanatory completeness is negligible in them.