Goldman presents a causal theory of knowledge designed to counter the Gettier cases that plagued the traditional view of knowledge (Justified True Belief). His theory effectively identified a problem with the traditional view, since in each case, the actor believed a fact, but the case was not the cause of his belief, which opened up the justification to be merely accidental that it applied to the circumstances. Despite conquering this fact in his theory, it cannot be a full account of knowledge because it has trouble reconciling inferences, cannot explain a priori knowledge, and does not account for knowledge not gained directly through a causal chain.
Goldman's causal theory of knowledge states that "an actor S knows that P if and only if the fact that P is causally connected in an appropriate way with S's believing P." The appropriate causal processes include the following: perception, memory, causal chains, and any combinations of the above mentioned. This definition entails that justification is causally constructed -therefore, how the belief forms determines the justification. An example of perception is our sense of sight. Through sight, we can know facts available through this appropriate causal process, i.e., that book is blue. This process is causal because the instance of light reaching our eyes causes us to see color. This causes our belief that the book is blue in our example. Our justification is how our belief was formed, via the fact that the light from the book reached our eyes. Therefore, this causal process satisfies on simplistic level Goldman's conditions for knowledge. The fact that P in this example is the color of the book, S believes this is the case be due to the causal process of perception to perceive this book, which caused the belief to begin with. An example with memory could be me remembering my girlfriend's birthday. Let's suppose I learned of this fact from my girlfriend telling me this fact. I believe her testimony was trustworthy, and that her telling me she was born on a certain date as well as my observation of her birth date on her driver's license. This makes my original knowledge of this fact founded upon a warranted inference, I believe her testimony is trustworthy and believe that her testimony must have resulted from the fact that she was indeed born on the date she claims. I might not remember when or where she told me, but my ability to remember ensures that I do indeed have knowledge of my girlfriend's birth date, and according to Goldman, it is because of the causal process of memory. An example of a causal chain being an appropriate knowledge producing process is an example that has been looked at by multiple philosophers is the lava example as Goldman states it:
"Suppose S perceives that there is solidified lava in various parts of the country-side. On the basis of this belief, plus various 'background' beliefs about the production of lava, S concludes that a nearby mountain erupted many centuries ago. Let us assume that this is a highly warranted inductive inference, one which gives S adequate evidence for believing that the mountain did erupt many centuries ago. Assuming this proposition is true, does S know it?" p.21
According to Goldman's analysis, there is a causal chain if in fact the mountain did erupt before S perceived the lava. In this case, the fact that there is lava was caused by the mountain erupting. The mountain erupting caused my perception of the lava. The connection between these two causes is an example of a causal chain. Causal chains remain causal even when inferences are added to them, provided the inferences are warranted, according to Goldman. These simpler examples demonstrate that the idea of causal chains is a relatively good idea when accounting for knowledge; however, the limitations of this theory will show that there is still work to be done to make this a complete theory of knowledge.
One of the problematic points that Goldman faces with his theory of knowledge is the contradiction between requiring a causal connection between the fact of the matter and the belief of the matter at hand and the use of non-causal processes to arrive at the belief. To clarify, whenever we formulate a belief from an inferential process that did not in a direct way originate from the fact of the matter, Goldman's theory will deny knowledge. An easy example is the following: suppose I form, the belief that I know steam was created in the process of creating the tea that I am drinking. Let us say my reasoning for this was because I know about how tea is made at this particular location. The inference was developed because of my observation of the current state of the tea. Goldman's theory and his explanation of it would work, if the cause of my inference was because of the fact of the matter that the tea was steaming. However, my inference instead resulted from an examination of the tea I had and the facts I knew about the production of tea. My illustration demonstrates that Goldman's account of inferences does not account for cases of a belief P with a justification containing inferences where the source of the inference is not resulting from the fact of P. Similarly, Goldman's theory does not account for a related process in justification: entailment. Justification seems to be preserved through entailment because for a given proposition, its entailment is a necessary and logical consequence. These logical chains of entailment are not causal, but arise as inferences. Seeing that inferences are not handled well by Goldman's theory, entailment does not seem to be in any better condition in examples where the justification ultimately resulting from entailment from a fact other than the fact that p results in a belief about p, but yet it can still be said that the person still has knowledge.
Goldman's theory focuses on empirical knowledge, which he states is the case at the beginning of his "causal theory of knowing" article. Goldman went as far as to state that the traditional analysis (Justified True Belief) was sufficient for accounting for knowledge of non-empirical truths. The problem with Goldman's stance on this matter is that there are not causal processes that link a priori truths to empirical truths, which would deny application of a priori knowledge to make empirical beliefs about the world count as knowledge. An example is the case of a smart mathematics student who spies the shadow of a house to be exactly 30 feet from the house (he knows because the shadow ends on a pre-marked area that is 30 feet from the house) that is 40 feet tall. He knows that top of the house is 50 feet from him through his mathematical knowledge of trigonometry. In this case, the knowledge has its justification in the entailment of mathematical knowledge, rather than though the causal processes that Goldman states are appropriate connections for knowledge. This case is important because the student is able to make an empirical belief based on a priori propositions in addition to information gathered through the causal processes (perception of the shadow, memory of the distance of the marker, entailment that the length of the shadow is equal to the distance of the marker). None of the causal chains however actually causes the belief. The origin of the belief is a separate a priori proposition that entailed that the given evidence necessitated the belief created. While there were some causal processes, I reject Goldman's view that inferences and a priori propositions when added to a causal chain make the entire chain causal. I reject this view because inferences and a priori propositions are not causal. If they were then inferences and a priori propositions would cause other causal processes. I want to say that a causal chain is causal until a link in that chain is not causal, any non-causal parts of a chain of reasoning cannot be counted as causal. A priori propositions in particular tend to either validate or invalidate one's justification for belief, but has no bearing on the actual truth of the situation. Going back to the house's shadow example, the trigonometric rules applied to form the belief had no bearing on the actual truth of the situation. The math student either validly deduced correctly the entailed answer given the input conditions, or does so invalidly. If the trigonometric proof is valid, then the justification is valid. The claim by Goldman is that the justification must be connected to the truth about the belief in an appropriate causal manner; inferences and a priori propositions do not constitute causal matter therefore according to Goldman's reasoning are not relevant in claiming whether a person has knowledge or not. Goldman attempted to dismiss this fact by reasoning that in all cases when an inference in a causal chain that the entire process that the entire chain is causal. Going back to my shadow example, while the belief is in reference to an empirical observation, that observation did not cause the belief, but rather the justification for the belief lies mostly with the mathematics that made it possible. To understand this, we can try to think if we take away parts of the justification. If the student did not have the a priori mathematical propositions would the student have been able to formulate the belief? If it is true that only causal processes can form beliefs about empirical observations, then knowledge is only possible through the causal processes; however, in this case, a non-causal process makes the belief possible. The observation itself could not justify the belief alone. One observing a shadow reaching a specific point does not cause the belief that the top of the building is a certain distance from the user; but rather this observation was utilized in the logical process of mathematics, and the mathematical process formulated the belief in the student. Therefore this process was not fully causal, but was also subjected to the a priori propositions of mathematics of which the conclusion became the belief.
Goldman's theory of knowledge covers a lot of ground and is a fairly effective account for many types of empirical knowledge, especially those with clear causal chains. However, as many theories of knowledge are, there are several conditions which can make a specific theory fail. In this case, one of the inconsistencies in how Goldman decided to handle inferences, a priori knowledge, and causal chains shows that his reasoning was either faulty or inconsistent such that the clarity of the case for knowledge is distorted. The reasons we can think of to demonstrate whether or not an individual has knowledge comes in many forms, and causal forms are only part of the processes needed for a person to obtain many types of knowledge. Having stronger support of inferential and a priori combinations of knowledge(in addition to causal processes) would lead to a better definition of knowledge, since it would account for the variety forms of justification we utilize in our day to day lives.