by Jim Pemberton
We often hear about how science is based in reason, but we aren’t often taught precisely how this works. We know something of the scientific method, but we don’t know how it relates to logic. We only have some sense that it does. We have come to the part of this series on Science and Faith where I will discuss how the scientific method is based on deductive logic in general.
There’s no need in this series to handle all the various forms that the scientific method takes, so this will only be a general discussion. It will involve the common steps and how each works together to produce a reasonable conclusion.
I had planned to discuss the limitations of the scientific method, indeed the process of scientific discovery, in this article as well. However, it occurred to me that there is simply too much material for one article, so I decided to split up the material.
The scientific method that you all learned in grade school is a good summary of it. The basic steps are as follows:
The word evidence is used for any of these, but evidence is a broad term that is part of both the Observation and Experiment categories. One thing that evidence isn’t is the conclusion. Evidence contributes to the conclusion, but it is not the conclusion. Many people refer to evidence as being proof of the conclusion. Evidence never leads to conclusion on its own and is never “proof” of anything. It is nevertheless important.
The conclusion is always the last thing. The reason I mention this in the preliminary remarks is to point out that many are tempted to start observations with a conclusion in mind. In fact, most scientific endeavors have a desired conclusion in mind. The good scientist is able to keep a desired conclusion from inadvertently skewing observations and experimentation. Not all scientists are good scientists.
We all experience the world around us. Given training in certain fields of science, scientists are trained to make educated observations in their fields. These observations follow a couple of patterns. One is calculated observations. For more mathematical sciences, this may involve simply sitting down with paper and pen and make formulaic theories. A scientist might also sit down and plan a structured observational event. A zoologist might plan a trip to the field to observe a certain type of animal and obtain data on its migratory habits for example. This leads to a more organic kind of observation. Something happens to be discovered and scientists are contacted to make trained observations on it or a scientist happens to be in the field gathering random data and notices a pattern (s)he wasn’t expecting.
The education of a scientist is important here. In making educated observations assumptions from the education, training, experience, and worldview of a scientist will be employed in order to categorize the observations accurately and effectively. These assumptions will be used throughout the process of following the scientific method.
Using the example of the light and the light switch from before, observation entails walking into a room and carefully looking at the things in the room. A window might be open, a chair turned to the side, the room illuminated from something other than the window, a rug on the floor, a switch on the wall, etc. A scientist might notice that the illumination of the room, aside from what is coming through the window, is coming from some devices on the ceiling. Some debate might ensue as to the source of the light from the devices on the ceiling. A careful observer will notice that the switch on the wall is able to be manipulated. When it is manipulated, the light from the devices on the ceiling changes its state.
From the observations a scientist makes, a sense of related occurrences can be noticed. From these occurrences, a relationship can be speculated. This relationship will be stated in a specific way. This specific way is called a hypothesis.
The hypothesis from our light switch example may be something like this: Whenever the light switch is up, the lights from the ceiling are on. This can be expressed in terms of a syllogism, although scientists don’t necessarily take the step to convert it to a syllogism. Since we have already used it as an example of a syllogism from last time, we know that the syllogism is something like this: If the light is on, then the light switch is up. As explained before, this is how a causal relationship fits into a conditional syllogism.
The scientist will then set up an experiment to test the hypothesis. That is to say that the logical correlation will be noted carefully. Not always will a scientist be able to manipulate one or both of the events, but as those events occur, their occurrences are noted for patterns of correlation. In setting up the experiment, a quantifiable measurement of success must be built into the experiment.
In our example, the scientist notes that the light switch can be manipulated and there are multiple devices on the ceiling, 4 to be precise. Success is determined when each device shows light whenever the switch is manipulated to the ‘up’ position. As a ‘control’, the windows are also monitored to see if the state of light coming from them changes when the switch is manipulated. The manipulation is set up to occur, say, 25 times. Any observation that the lights are on when the switch is not up are also catalogued.
The experiment is conducted as planned and the results noted for each light in the ceiling and for each window. These results are analyzed using the criteria established from the beginning. So what comes up is generally a percentage representing a likelihood that the hypothesis is true. Very low percentages are typically deemed to be proof that the hypothesis is false. Very high percentages are typically deemed to be proof that the hypothesis is true.
In our example, we have no observation that any of the lights are on when the light switch is down. We have no observation that the light from the windows changes. However, we observed that each time the light switch was manipulated to the “up” position, the lights in the ceiling usually came on. (Interestingly, it was noticed at this time that when the light switch was manipulated to the “down” position the lights in the ceiling went off. This was not part of our original experiment, so we can’t count it, but we can include it in a future experiment.) The exception that occurred is that after 11 times, one of the lights stopped turning on. So there are 14 counts of the light switch not affecting one of the lights. The conclusion therefore is that we have 86% likelihood that the light switch affects the lights.
In order to be seen as acceptable science in the scientific academy, two or three other things must happen. It should be published with peer review, and in accordance with continuing peer review, it must be confirmed by another group of scientists repeating the experiment to verify the same results.
Now this was a ridiculously simple experiment. There were some key concepts that it didn’t cover. Nevertheless, we covered the basic nuts and bolts such that what I hope is understood is that the scientific method follows the pattern of deductive logic, that the testing results in a likelihood that the deductive syllogism is set up.
In the next article, I will discuss some of the limitations of scientific discovery.