Agnosticism, Atheism, Christianity, Faith, God, Religion, Truth

Difficult Questions for Atheists? Part 1

A little while back, my friend UnkleE suggested that I should consider some questions that he believes are problematic for atheism as a worldview. He listed 5 questions, and I want to take them one at a time, so they can each get the focus they deserve.

That said, my initial responses to each of these questions may not be very long. Instead, I’d like to use each of these posts as a launchpad for discussion. I know these are issues that UnkleE (and probably many of you) have thought about at length, and I’d like to consider those arguments as fully as possible without subjecting everyone to my own rambling preamble. So, here’s question 1:

Do we have free will? If so, how? If not how can any choice be based on evidence rather than brain processes?

I don’t know.

I’m aware that a number of physicists and other scientists sometimes argue that free will is an illusion. That was shocking to me when I first heard it, but I now realize what they’re saying.

Imagine you could go back in time to a point where a decision was made on something seemingly insignificant. In 1959, Buddy Holly, Ritchie Valens, and the Big Bopper died in a plane crash. It’s said that another musician, Tommy Allsup, was going to be on the plane instead of Ritchie Valens, but they flipped a coin for it, and Valens “won.” If you could go back in time and witness that coin flip without interfering, would anything happen differently? Whoever came up with the idea of flipping for it thought of that for very specific reasons that would still be the same if it happened over again. Allsup flipped the coin at a specific level of force, and it flipped through specific atmospheric conditions. Those things would still be the same if you were watching it happen. Valens called “heads,” which he did for specific reasons, even if they were subconscious. In other words, every single thing that happened, even though they were seemingly random, happened in particular ways for particular reasons. If you could replay it over, there’s no reason to think anything would play out differently.

And every decision you’ve ever made, you made for specific reasons, even if the decision was close. If you went back in time and made the decision over again, but only knew the same things you knew at that moment, could you have made any other decision?

There’s no real way to test this, but the thought experiment leads many to conclude that true “free will” is not really possible.

I’m not sure how I feel about that. I do think that if you could replay decisions, it’s unlikely they would ever change. But that’s not really what I think of when I think of free will. Just because I made all my decisions for specific reasons and was “powerless,” in a way, to do anything different, that doesn’t mean that I had no control over the decisions. Thought processes were still firing in my brain as I calculated a number of factors, considered past experiences, estimated probabilities, and tried to predict possible outcomes. I might always come to the same conclusion in the same circumstances, but my mind is still very active in the process.

[H]ow can any choice be based on evidence rather than brain processes?

I think any choice — any good choice — should be using both. Brain processes deal with information, and that’s all that evidence is, so I see them as being very closely related.

Ultimately, I don’t see how this question causes a problem for atheism. I may have more to say about it in the comment thread, but I’ll need to see the case against atheism filled out a bit more before I can really weigh in on it.

322 thoughts on “Difficult Questions for Atheists? Part 1”

  1. Ark: I understand the motivation behind your comment, and your long history with this blog (which I don’t have). But actually, I was never “un-gratious” towards unkleE, I would say quite the opposite. If you remember, it was not me who had the tantrums. In fact, I took his comments seriously and spent time attempting to address just about everything he said very carefully and directly.

    I get the impression that with folks like unkleE, if you express blunt disagreement or if he gets the sense that he’s unlikely to convince you of his own views, or worse, that his arguments have been deflated, these deflated arguments suddenly become “dead horses that needn’t be flogged further.” Apparently, ad nauseam repetition and ignoring objections, are supposed to be lethal to stampeding horses. At this point, he loses interest, and he informs you that you have been bestowed the “unkleE cold shoulder award,” as nothing that you could possibly have to say is worth further consideration (not that much was ever considered in the first place), and that henceforth you shall be ignored, to your everlasting shame. All of this is in the hopes that you’ll go away, so he can continue to try to convince others. This is simply a manifestation of advocacy for a preconceived notion, not of engaging arguments on their merits with any kind of an open mind.

    I’m still here, although not wanting to disrupt Travis’ discussions, as I’m curious to see if Travis will manage to break free of the two-step shuffle.

    Liked by 1 person

  2. Speaking of the two-step shuffle, so far this is what we’ve seen offered:

    1) There are questions that “atheism/naturalism/physicalism/determinism” haven’t resolved, therefore there are “problems” with those views. But never mind whether the opposite views “supernaturalism, dualism, etc.” or any other views can solve anything, because the onus is always on “atheism/naturalism/physicalism/determinism” to answer every question, and if they don’t, then we can talk about whether there are problems for any other views, because those are separate issues, you see. (Meaning: if your “worldview” can’t solve these problems, never mind that mine can’t either, because mine wins by default.)

    2) Cause-effect processes cannot produce ground-consequence cognition via Natural Selection. (Meaning: cognition can’t exist without libertarian “transcendence,” period, unless designed using libertarian “transcendence”).

    3) “You gloss over the problems of cause and effect processes producing ground and consequence reasoning via natural selection, without explaining them in enough detail to show they can reliably deliver…” (Meaning: if you don’t dot every “i” and cross every “t,” why, Natural Selection is a total non-starter; never mind that goddidit explains even less.)

    4) Natural Selection could not possibly select for cognition. Well, maybe it can select for some cognition, but “higher cognition”? No way, that’s out of bounds, because NS can only possibly select for jerky reactions, like flight or flight. Because, you see, if you stop to sit there and “think” in a predator situation, you might die, so therefore intelligence and cognition beyond herky-jerky reactions must be completely useless for survival and reproduction. (Of course, no explanation is given for this to counter all the evidence to the contrary). (Meaning: never mind that there are many other situations–besides ones where you have to react quickly–where cognition could be selected for; never mind that general processors with plastic intelligence are demonstrably useful for reproductive fitness, never mind sexual selection, peacock tails, evolutionary by-products, spandrels, and all those other concepts that evolutionary biologists and psychologists talk about at length. All those can be dismissed with a sweep of the hand, a straw man, and a fallacious appeal to “incredulity.”)

    5) It is “telling” that (Travis and Dennett) bring up the example that logic, by way of physical processes, works for computers. But that’s because they’re deliberately designed. (Meaning: cognition can exist without libertarian “transcendence,” but only when designed using libertarian “transcendence”).

    6) Grounds never produce consequences on determinism, or not “reliably” enough. (Meaning: cognition can’t exist without libertarian “transcendence,” period.)

    7) Sure, a computer can do logic through cause-effect processes, provided it’s been designed bug-free. But it’s only doing what it’s been designed to do. (Meaning: cognition can exist without libertarian “transcendence,” but that’s because it’s designed using libertarian “transcendence.”)

    8) Assumption of “transcendence”? No, of course not, no assumptions at all, none! “Just saying” that cognition is not easy to explain on Determinism. (Meaning: libertarian “transcendence” is not being assumed, but if it is not assumed there are “problems” that are difficult to explain; never mind that assuming it doesn’t solve any “problems.”)

    9) If logic doesn’t transcend the physical world, then either: Possibility 1: Physical processes can execute logic, or Possibility 2: Unless there is libertarian “transcendence,” there would be no logic. Determinists blindly accept Possibility 1, but that fails because they haven’t shown the mechanism, and we should really prefer Possibility 2. (Meaning: Logic <strong<can't exist without libertarian “transcendence” unless the mechanism that gives rise to it can be explained in excruciating detail. Never mind computers and Natural Selection, those don’t count, they’re all a rationalization by Determinists. And never mind that Nondeterminists have offered no mechanism whatsoever, they’re just “saying there are problems.’)

    10) Predictions can be random, if random exists. But when it comes to human prediction it can only be ground-consequence based. If it’s cause-effect, it’s only a herky-jerky reaction, like “flinching,” because you see, NS could only ever select for very quick things like “fight or flight,” there would never be any time to think, and it would only be an “unconscious reaction.” (Meaning: beyond herky-jerky unconscious reactions, pensive predictions can’t exist. because Natural Selection cannot possibly select for anything outside of where speed of decision is required.)

    Personally, I’m still waiting for any of this of this to make any sense, or to have any justification behind it.

    Liked by 1 person

  3. But actually, I was never “un-gratious” towards unkleE,

    Never once suggested you were anything but the gentleman and the scholar, AP. I think you’re handling your ”end’ very well under the circumstances.
    I merely suggested that you are now another who has been the brunt of the notorious brand of unklee’s style of condescension.( and if it were just me saying he behaves in this manner one could easily say: Oh, Ark is just having a tantrum.”
    But ask people like Ken (KCChief), Gary, William, Scottie Nan, Carmen, Violet … and now You, etc. I am sure you get the picture?
    I often refer to his dialogue over Nazareth on his blog with a chap called Bernard. It is the perfect ”primer” if you wish to understand exactly what you will be dealing with when engaging unklee.
    He don’t think he likes me very much. I upset his didgerydoo-style apologetics once too often. But that’s okay. Can’t like everyone I guess?
    I am sure there is probably a lot on nice things about Eric if you dig a little, but if I have to dig for goodness I prefer doing so in the veggie patch in my garden!

    I shall of course be following along with keen interest, a certain amount of malevolent glee – ’tis my nature after all – and a fair dollop of amusement!

    Have fun.
    Ark.

    🙂

    Liked by 1 person

  4. Eric,
    I haven’t found the time to put together the kind of response I want to offer and I’m heading out to the woods again for the next few days, so for now I’m going to redirect you to something that I think can set the stage for further discussion. Just a couple weeks ago the following article was posted and I think it does a pretty good job of summarizing a model that I find particularly promising:
    View at Medium.com

    This will help explain why I asked the question about prediction. For the time being, I would be interested just in your response as to whether you find that this model is a reasonable evolutionary product of deterministic physical processes and, if so, whether you recognize an opportunity for it to scale into a system that can respond to increasingly complex situations.

    I’ll be back Sunday night to continue from there.

    Like

  5. Hi AR,

    I have been pondering whether and how to respond further following your comments on our discussion. I don’t like to be offside with people, so I thought I would try to build a bridge. Hopefully I won’t simply deepen the ravine between us.

    You have writted how you have seen it, so I will briefly offer you my perspective.

    We are all free to comment, or not. I don’t believe I am under any moral obligation to respond or to reply in the way someone else wants, and I don’t place any obligation on anyone else. I try to politely respond to comments addressed to me, but I can’t respond to everything people say. Discussions cannot go on forever, so if I think things are no longer constructive, I stop, and I’m not under any obligation to go on – the obligation I feel is to be honest and courteous.

    The internet makes us all fairly anonymous, and it is easy to say things we wouldn’t say face-to-face. I try to run all I say through that filter, and when I occasionally slip up, I apologise. I don’t regard these discussions as a competition, so if someone is misunderstanding me, I say so, simply to clarify, not to blame or score points.

    These views may help explain how I have approached our discussions. I’m sorry my approach didn’t please you, and I hope there may be no bad feeling between us.

    Like

  6. Hey Travis, that’s a fascinating article, and from an interesting source that I wasn’t aware of – I have signed up. So even if nothing else, I will have profited from this discussion! I’m not 100% sure I understood it all, but I think it may well be a useful model. However I’m not sure how it can deliver what you want, but I’m happy to learn.

    I have read a little about prediction since you asked me the questions, and I’ll be interested to see how you put it all together. Thanks. (I’m quite happy to wait, it gives my brain some time to rest! Enjoy your time in the woods. Are they literal woods, are you bushwalking, hunting, felling trees or preserving them, or …. ?)

    Like

  7. unkleE: We agree that there’s absolutely no obligation to respond and that there needs to be courtesy. My approach is to stick with arguments, issues, etc. and not persons. I think statements are fair game to be criticized as long as it’s not made personal. Also, I enjoy the exchange and I think it forces me to think and clarify my own positions and understand them better. This, along with someone else’s perspective, can only lead to learning, which is the point for me. If my bluntness and directness comes across as personal, it is not intended as such and I apologize for that as well. Cheers.

    Like

  8. Travis: Nice article. The concepts are not new from the Machine Learning perspective, but the authors’ key insight that our stream of consciousness can be modeled as a continuous prediction of an “expected” world punctuated by surprises coming from new information is very interesting.

    I think one key aspect of ML that may be helpful to understanding biological cognition (including, possibly, decision making) is how these machines are not programmed to do specific tasks, but are given a general-purpose architecture with random initial states, and are then allowed to learn from examples instead of from hard-coded rules. No two of them ends up behaving exactly the same way, and their reaction to information that they’ve never “seen” before leads to surprising positive results even for their builders.

    I’ve been teaching a graduate-level course in ML for the past 5 years, and my students have come up with very interesting final projects. Last year one team replicated a study of an ANN-based classifier to classify 15-second snippets of music by genre (rock, classical, jazz, etc.). Individual classifier accuracy was in the low 90s% when the networks were presented with music waveforms that they’d never encountered before. No two networks were exactly the same and there was “disagreement” in some cases. Yet, when combining several of them together to “vote” increased their accuracy to the high 90s%, which is reminiscent of “expert committees.”

    Like

  9. Interesting discussion! I have a question for UnkleE:

    If we define the natural world as Realm A and the supernatural world as Realm B where Realm A is non-determined due to interactions from Realm B: Will the combination of Realm A and Realm B be a determined or non-determined system?

    My guess is that the answer to this depends on whether you believe Realm B is a determined or non-determined system. If it is determined then the combination of both systems will be determined. If it is non-determined then it contains some factor of randomness or it contains interactions coming from a third system (Realm C?).

    The point I am trying to make is that determinism is ultimately the only option available unless you believe that true randomness is possible. So we can either accept determinism or we can try to prove how true randomness is possible. Or perhaps I am missing something.

    Like

  10. Hi Dave,

    This discussion is about difficult questions for atheists, i.e. asking how an atheist responds to determinism. What you are asking is a different topic, and I think it best not to divert from it here.

    But my brief view is that you have assumed, I think, that either things are determined or they are random. But perhaps there are other types of actions, ones not fully determined but in some senses freely chosen after consideration of the issues. Like in fact we all (I think) feel we experience all the time.

    Like

  11. This discussion is about difficult questions for atheists, i.e. asking how an atheist responds to determinism.

    I accept determinism since that is what appears to be the case: cause and effect. I will also accept randomness if that is ever proven to exist. Your third option, free choice, does not make any sense to me. I don’t see how a choice that originates from an uncaused cause would be different than random.

    Like

  12. nate,

    are you ever going to share the 2nd question for atheists?

    This current one seems a bit too similar to discussing what non-carbon based life may look like – really just guesses and conjecture, with little to no support. I still don’t see the problem for atheists, nor how theists wouldn’t have the same dilemma…

    Liked by 1 person

  13. Hi Dave,

    “free choice, does not make any sense to me. I don’t see how a choice that originates from an uncaused cause would be different than random.”

    I think you experience choice every day, multiple times.

    For example, did you make this reply to me because you thought it was true, or because your brain was determined to? If because you thought it was true, then perhaps you can join Travis in explaining how a physically determined process can produce ground-consequence logic. But if your reply was merely the result of determined physical processes, why would you think it would be meaningful or true?

    I don’t wish to be rude either to you, or to William, but I think the reason why few seem to have a problem with determinism is because you seem to be happy to accept determinism without an explanation for how we actually live our lives every day.

    Like

  14. Ark, “But ask people like Ken (KCChief), Gary, William, Scottie Nan, Carmen, Violet … and now You, etc. I am sure you get the picture?”

    Sorry it took me so long to see this, Ark. The last 100 comments on determinism have passed me by. I don’t mind dipping my toe in a conversation such as this, but these guys are using diving bells . 🙂
    But to comment on your statement above, you are correct to assume that I feel unkleE comes across condescending in many instances whether he means to or not. I have also noticed that once he feels he has achieved the upper hand OR that you have trapped him, he simply ignores you going forward. I have given unkleE no reason to ignore me as my comments have been civil and yet about 6 months ago he decided he didn’t need to waist his time addressing my questions.

    Like

  15. Eric,
    Sorry I’ve been slow to respond. Life gets in the way. And yes, I was in literal woods. I live in an area with easy access to great hiking, camping, etc…

    I suspect that this discussion will come down to unpacking your reasons for suspecting that a predictive model cannot yield a process for making generally accurate logical inferences, but I’ll try to expand on the article to establish a link to this reasoning thing that we do.

    Let’s start by talking about concepts – those mental representations of things, adjectives, events, actions, etc… The neural model proposed in the article might say that a concept is the pattern formed by the deeper layers filtered down to capture the “surprising” or distinct aspects of the inputs. So concepts are discrete in the sense that they are a distinct pattern or set of pattern associations. This coincides with our disposition toward essentialism and is foundational to the system of logic, where we treat everything as discrete entities that possess properties or characteristics (associations). Most formal logical arguments make this assumption even though the essentialist paradigm is usually a fallacy that ignores the extent to which entities actually lie on a spectrum. If this model is correct, our neural architecture is emphasizing the differences between sensory inputs and this contributes to the way we partition the world.

    Regardless, this sets the stage for logical evaluation of the relations between these entities. Recall that the learning which builds these patterns and associations is generally based on the principle that “neurons that fire together, wire together”. This learning is a type of statistical inference that is ultimately driven by real-world inputs from sensory data. So the patterns and associations are reinforced by repeated exposure to the regularity of the world. When we couple this with the predictive model we end up with a system capable of not only accurately capturing the statistical associations between inputs, but accurately predicting the consequences of those inputs, as informed by the associations established in prior experience. The favoring of the most probable prediction comes as a natural consequence of the learning that configures the predictive coding network.

    Let’s try this out with a classic example of logical deduction: “All men are mortal. Socrates is a man. Therefore Socrates is mortal.” If we assume that our model comes to this syllogism with a relatively blank slate, only previously containing the concepts of men and mortality, then the model would not presuppose that the new entity “Socrates” has any established association with mankind or mortality. The conclusion will be a completely novel prediction. Upon receiving the semantic content of P1 (All men are mortal) the network creates a new association between the existing concepts of men and mortality. Upon receiving the semantic content of P2 (Socrates is a man) the model has created a new concept for Socrates and a corresponding association to the concept of men. The net result (no pun intended) is a new association between Socrates and mortality by way of the associations of the intermediate concepts as they were introduced into the model. If we are to then prime this model with the semantic content of “Therefore Socrates is…” the predictive paradigm would build upon the fresh association to predict the mortality of Socrates via the association. If, however, the model previously learned an association that ran counter to a premise then the statistical strength of the prior association would override the proposed new association and lead to a prediction in opposition to the proposed conclusion. Of course, this is all very abstract and glosses over the physical cause-and-effect activity which underpins the process, but I hope that the article and prior discussion helps clarify those aspects. And though this may be a relatively simple example, I struggle to see how it does not feasibly scale to support more complex predictions.

    Lastly, I should also note that there is a relatively new and popular theory of reasoning’s origins called the “Argumentative Theory of Reasoning” which is largely motivated as an explanation for cognitive biases and partially motivated by the idea that it makes no sense for us to have evolved introspective reasoning as a mechanism to support or correct our beliefs because we generally just trust ourselves. In this theory, then, reasoning evolved not to pursue truth but rather as a social tool for motivating convergence between members of the tribe. The trick though, is that reasoning can only be effective in this way if we also have evaluative faculties that can be persuaded by the reasoning of an argument. The progenitors of this theory suggest that reasoning as argument and reasoning as evaluation co-evolved. I am inclined to believe that this theory offers a good explanation for the strength and prevalence of some biases but that it needs a foundation on which to operate. I would suggest that the multi-layered, recursive statistical inference can serve as that foundation, so that the use of language and the accompanying social pressures in the argumentative theory are essentially catalysts which accelerate the development of our capacity for reason, encouraging more complex and abstract forms of reason that still utilizes a foundation of statistical inference. In that case, the associations are still probabilistic and driven by real-world inputs and so will generally reflect the way the world actually is – meaning that the convergence achieved through reasoning as social interaction can still operate in the direction of truth.

    I’m not an expert in machine learning, cognitive neuroscience or evolutionary biology, so I’m sure there are flaws in my presentation, but that’s how I currently see it. What do you think?

    Liked by 1 person

  16. AR,
    I agree that the generalization inherent in these learning networks is a key concept to understand. If you’re teaching graduate courses in machine learning then you certainly are more informed on the topic than I am, so feel free to chime in with any corrections or additions as you see fit.

    Like

  17. @ Kcchief.

    …and yet about 6 months ago he decided he didn’t need to waist his time addressing my questions.

    my emphasis

    Perhaps unklee eats a lot of burgers these days and he will get round to answering, soon?

    Sorry, Ken, it was such a good typo just sitting there begging that I couldn’t resist.
    🙂

    Like

  18. Unklee said:

    I don’t wish to be rude either to you, or to William, but I think the reason why few seem to have a problem with determinism is because you seem to be happy to accept determinism without an explanation for how we actually live our lives every day.

    How, or rather em>why we do the things we do?

    Either way if we are merely speculating, then it really is only a philosophical exercise for atheists. In truth most do not give a monkey’s unkle. ( if you’ll excuse the pun, here)

    However, for the theist … that’s you, of course, once yet another way is found of removing the god-factor – or at least demonstrating again the spurious nature of your virgin birth god-claims – and to be precise here, we mean, of course your god, the make- believe character, Yahweh, either au natural or in his Jesus disguise, you are proverbially screwed.

    Is this why you try so, very very hard to prove your points all the time?

    Liked by 1 person

  19. “I don’t wish to be rude either to you, or to William, but I think the reason why few seem to have a problem with determinism is because you seem to be happy to accept determinism without an explanation for how we actually live our lives every day.” – UnkleE

    UnkleE, I haven’t found you to be rude, so you and I are good there.

    But I don’t know if “happy” is the word… “confused” is probably better.

    This confusion is probably my fault, as this is a conversation I’ve largely avoided – not because of the difficulty I think it places on my views, but because I think it’s entirely moot, and seems too much like trying to make a mountain out of an imaginary mole hill.

    I make decisions and choices based on a variety of factors: my emotional state, my knowledge, my ignorance, my sense of right, my sense of selfish wants, short-sighted-ness vs foresight, etc… I fail to see where these present any problem whatsoever, and I fail to see how believing in god makes them different.

    How does one make a decision or choose something any other way? Can someone really decide what they think is true based on nothing?

    The silliness of the thought reminds me that I haven’t been following the discussion and very likely have missed the key part that shows how this is even supposed to be a problem… but it’s too rooted in supposition and imagination to seem valuable to me. Sure it’s fun to toss around in the backyard with friends, but that’s a game or idle exercise, and not a Superbowl bound together by tangible rules – so how does one judge who really wins, who’s really right?

    so if everyone wants to keep digging this pointless hole, I’ll show myself back out, but I am still curious to see some other questions that may be problematic to my view point.

    Liked by 1 person

  20. unkleE: I don’t wish to be rude either to you, or to William, but I think the reason why few seem to have a problem with determinism is because you seem to be happy to accept determinism without an explanation for how we actually live our lives every day.

    The reason that I don’t have a problem with determinism, is that I see no evidence that the cosmos is deterministic. It would be foolish of me to have a problem with a thesis that is likely false.

    Liked by 2 people

  21. For example, did you make this reply to me because you thought it was true, or because your brain was determined to?

    Both. It was determined based on all of the complex factors that our brains sift through while making decisions. All of our prior knowledge, preferences, mood, character traits, instincts, etc. It is a choice that is determined based on a very complex neural system that we have no way of self-predicting. (Interesting side note: It has been reported that scientists using brain imaging techniques can “see” simple choices before people are aware of them. https://www.google.com/#q=scientists+see+your+choices+before+you+are+aware – if this is true I think it would support determinism and LFW would be less probable.)

    Why do I think something I’ve written is true if everything is determined?
    Because we have a neural system that relies on a vast amount of stored information, repeatable and predictable results, and has the ability to filter out ideas that are not aligned with observable reality. The same cannot be said for a choice that originates from an uncaused cause. Why would you trust this type of choice? How is a choice that originates from an uncaused cause different from a random event?

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s