If you watch my front door long enough, you will soon notice a pattern: I get home from work about the same time everyday. If you didn't know better, you would be tempted to believe there is a "law of the universe" which forces me to be home at that time. But you know there is no such law, that the time I get home is the result of a personal choice. The fact that there is regularity and predictability to it doesn't imply any causal explanation. In a sense, I get home everyday at 6:20PM for no other reason than because I want it so.
This is a popular antirealist argument and have mentioned it on this forum from time to time. Let's say for argument's sake that all events in the universe are such as your situation about arriving at home at 6:20PM. We never catch a mistake in this process because we either haven't sampled enough early or late arrivals to see that the universe is not so lawful, or the universe just does what it does for no reason and it just happens to arrive in a manner which we deem as lawful.
In the latter case, if we say things happen for no reason (no cause), then we are stating that the world is random, and in which case there is no point in discussing the situation since everything I would say to counter would only be construed as attributing patterns to random interactions in the world. A meaningless debate.
In the first case, we are addressing the reason the possibility that you could arrive late or early depending on traffic, if you work late, etc, but we just haven't stood in front of your house long enough to see the potential variability in your arrival. Had we done so, we would see right away that our theory on your arrivals is not so accurate but is approximate. Okay, so far so good as to what we see in the laws of physics. We do more observations, and finally we recognize that our former theory only approximates the phenomena and we need a better theory to explain the exceptions in a manner consistent with the former theory. But, notice that the new theory has to explain not only why it is that you arrive late and why you arrive early on some occasions, it must also explain why it is you arrive mostly at 6:20PM. In other words, the theory is expanding its reach not only to what we notice on a day in and day out observation, but it must now consider what happens outside of the scene of your daily arrivals at home. We must be concerned with what the situation is like at work, how is traffic, what you are thinking that day and how that affects your home arrival, etc. This is what science does, and it still is able to predict the exceptions as rare as they are on occasion (even if it is in probabilistic fashion). Is it inerrant? No. There is always the potential for another variable to exist outside of the ones we have considered which detrimentally effects our theory. So, your argument sounds like it holds up.
However, something strange begins to happen as we expand our theory to encompass what originally was focused on what happened outside of your house at 6:20PM to the wider scene as to what probabilistic issues might affect your arrival. Our theory becomes more and more simplied and unified with other theories that we had of not just your arrival, but your neighbors arrival, and arrival times for every citizen of Toronto. The individual theories become unified such that we can make a whole series of predictions that work well in your situation and a whole bunch of other people.
Now, why would that be so if we were only tracking what was happening in some happenstance regulative manner? Why would we obtain a deeper level of explanation of something that gave us a whole lot more explanative and predictive power than we had previously unless we were actually reaching the facts of the matter in some asymptotic process? Think about this for a second. How do we know when we have grasped something basic such as when we have just ate a hamburger? We do so because everything around us is such that if we didn't just eat a hamburger, then nothing that we ever did would have any meaning. In essence, if we didn't just eat a hamburger, then we should consider nothing as true - not even our own existence.
Yet, if we begin to doubt the same reasoning that led us to develop a very good understanding and predictive abilities as to why Toronto citizens tend to arrive at their homes as predicted, then we should also doubt the same reasoning processes that led us to believe we just ate a hamburger, and hence the very things that lead us to believe that we have done anything or that we even exist. However, this is meaningless since to even think about such things we are already assuming our existence as separate beings (etc), and the very act of engaging in thought is to assume some things as true - including our ability to reason and obtain factual information about the world.
Could it be wrong? Of course. But, if it is wrong, then to even entertain what is correct or instrumental or empirically constructive is meaningless. We have to grasp onto the world as a real entity which we can for all practical purposes understand, and we just accept that as a given.
1) Dismiss the issue completely. Thanks to science we can do a lot of useful things, so who cares?
This is not a valid choice since, as I have shown, you cannot dismiss the ontology of science without dismissing the ontology of our experience, and hence the sensibility of our choices. If we cannot make sensible choices based on reality as it appears as factual to us, then this also applies to 'useful things'. Useful things are only as good as they are meaningful to us, and in order to find meaning in anything we must apply those attempts for meaning in a manner consistent with our experiences (e.g., logic, math, science).
2) Face the issue from the only perspective we trust as reliable: logic.
If logic demonstrates that there is no real dichotomy between scientific ontology and human meaning then we cannot so easily confine science into a neat little corner that we label "useless as telling us about the world as it is, but helpful in engineering issues".
It's the road Dick pursued, and I'm not qualified to tell whether he suceeded or not. However, I do take the position that until the issue of science and truth is properly addressed, I have no intellectual obligation to accept science as true. Some people think I'm anti-science, but I'm just trying to be consistent.
I don't think you have an intellectual obligation to accept science as true, I think you have an intellectual obligation to stay consistent with any approach that labels science as unrelated to truth. This is the pit that antirealists fall. It is not possible to remain in any way consistent to an anti-realist stance and still look with a straight face that sunrise is at a certain time, that the weather will probably be cloudy tomorrow, etc. Once someone dismisses scientific truths, they are in a position, I think, to dismiss all science as truths, and then they must explain to the rest of us why they are carrying an umbrella when the weather forecasters have predicted rain that day. Are we to believe that they do not believe it will rain, but just see that information as useful? Those kind of answers are ridiculous in my view.