Question 1 Hempel’s D-N model outlines the way an explanation should be formulated and stipulates the characteristics it must have. In short, statement S is an explanation if S contains both an explanans and explanandum. An explanation is, in essence, an argument, with laws and conditions making up the explanans and the explanandum being reached via the explanans. The explanans must meet three criteria and the explanandum must meet one criterion. If these criteria are met, says Hempel, the explanation is ‘good’. The first criterion of the explanans is that it must contain one general law. A general law is a true universal generalization. It is not merely a generalization that is “accidentally true”. An accidentally true generalization such as “all the people that walk into this classroom are students or professors” does not offer any real explanation and, therefore, is not a law. The second parameter for the explanans is that it must contain empirical content. This idea is rather complex, but it generally means that the statements of the explanans are not simply true by nature of definition. Using the “S is a bachelor” argument as an explanation to the question “Why isn’t S married?” offers no explanation as to why he is not married, it simply classifies him by a property that is built in to the definition of “bachelor.” While this generalization is not accidental, it contains nothing that actually accounts for the fact that he is unmarried. The final criterion that the explanans must conform to is the property of trueness. The statements within the explanans must be true (i.e. “corresponding to how things actually are”). Hempel says that an explanation is not correct unless it reaches a true conclusion through true statements. The Explanandum has only one criterion, and that is it must follow logically from the statements presented in the explanans. Essentially, the consequence must be validly deduced from the arguments. This argument, basically iff S then D where S is the explanans and D is the explanandum. This set of criteria and the way that Hempel states that they must be formulated is rather powerful and can validate quite a few explanations while invalidating many false ones. However, the D-N model is not flawless. The truth requirement, for one, is problematical simply because there are too many things that cannot be known for certain. Countless theories (some of the bigger theories including historical heat-energy theory and “the earth is the center of the solar system/universe”) have been widely thought to be true and have later been falsified. The problem, then, is in many cases there is no way to know if some statement, X, is actually true. The statement “hydrogen has one electron” could very well be false (and proved so) in a year or 100 years. This then retroactively invalidates the explanation, whether or not it has been “useful” for the time it has been in use. A further problem with the D-N model is that there are many (infinite?) effects that are caused by the same thing. In these cases, Hempel’s model can be adhered to while using one of the effects to “explain” the other effect. This happens when there is some common factor between the two effects and the cause is hidden or unnoticed (recall the barometer example). The hidden cause
creates both effects so neither effect is explained by the other, even though an explanation can be formulated that way. Another complication with the D-N model is that there are some cases where the explanans and explanandum (and the statements making up each) are in a way symmetrical so that the argument can actually be reversed (recall the flagpole example). The actual explanation is not symmetric in the way the argument is, however, thus creating an invalid explanation that fits Hempel’s model. These flaws in Hempel’s model are quite severe, showing that in some cases the criteria are not strict enough. From what I have read, it seems that many philosophers feel this could be remedied with the addition of some sort of causal criteria or concept, requiring that some sort of causal link be present in an explanation. Kitcher’s unificationist idea is a different way to render the D-N model more accurate. Kitcher says that things are explained by fitting them into generalizations or patterns, such as Hempel’s D-N model. Kitcher urges us to realize that not only do we do this with individual things but also with our total set of explanations. He says that the more things you can explain using the fewest general patterns, the better your total set of explanations is. This applies to explanation in a very similar way as Occam’s razor. Given two explanations of the same occurrence, the one that relies on our existing set of explanations the most, or causes us to adopt the fewest new explanations into our set (i.e. is the simplest), is the correct one. The basic idea is that reducing the total explanations is a way to unify them (hence unificationist theory). Extrapolating the effects of this theory, one could easily hold a very incorrect set of explanations for a period of time, but in the end (given enough time) the total set of explanations would be changed and altered to fit truth (i.e. how things actually are). While this seems like a foolproof way to get to the truth and correct explanation, it’s almost like trial and error in the sense that you may only reach the right explanations by using every possible wrong explanation. However, at this time (in my philosophy experience) unificationist theory is the only theory that I can see ever reaching pure truth and therefore wholly correct explanation. Question 2 Observational terms and theoretical terms are quite distinct, just as their material (or potentially material) are (observational and theoretical things). Observational terms characterize that which we can observe (not just visually, but sight is a very important aspect), or experience. Terms that describe shape, color, and other definable/measurable things are meant as observational terms. Theoretical terms are terms describing that which we cannot directly observe; for entities that are assumed to exist to explain the things we can observe. Observations (and therefore observational terms) are key to justifying theories, comparing theories, and giving meaning to theoretical terms through bridge/correspondence laws. However, there are several problems with these definitions and the distinction between the observational and the theoretical. The first stems from the definition of measurable/observable. Surely one would agree that using a microscope does not taint the observability of an entity. But the same may not be so apparent
when speaking of devices of measurement, say thermometers, voltmeters, and so on. In fact, a voltmeter can be said to measure properties of the theoretical flow of theoretical objects (electrons). In this case, it is not really an observational tool at all. The same type of conjecture can be applied to other devices such as thermometers. A further problem arises when we consider that a theoretical entity at one point in time may later (due to improved technology, a change in understanding, etc) be considered an observable entity. Putnam discusses the implications of this in his work: when a theoretical entity becomes and observable entity, many of the terms used to describe it can still be applied, thus blurring the lines between observation and theoretical terms. As Putnam states, many (if not all) theoretical terms can be equally applicable to observable entities. Putnam also shows that the status of being “unobservable” does not mean that an entity is theoretical; the opposite is also shown: that observability does not exclude theoretical properties from an entity. A better way of defining an observable term/property may be as stated by Professor Richard Feldman of the University of Rochester: Property P is an observable property of object X for observer S at time T iff at T, S is able to determine by observation (as opposed to theoretical inference) whether or not X has P. Hanson, however, would not agree with this description. Hanson believes that observation is inherently intertwined with theory. He does not believe that “S is able to determine by observation” without theoretical inference, whether it is conscious or not. Hanson’s argument is that we are incapable of making a pure observation because of the experiences we have and the knowledge we hold. Essentially, he is saying that in the act of observing, what we observe is innately based on knowledge, experience, and contextual information. While inference can be active and strenuous (not everything is obvious at the moment of observation), to Hanson there are many different things affecting inference at many different time intervals, from instantaneously to long after the observation. The main point of Hanson’s being that the theory, or what we intend to – or think we should – see, heavily influences what we actually observe.