Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Knowledge of everything

  1. Oct 23, 2005 #1
    Is it possible to know everything?

    If it is possible, then is it dependent on time? that is, for a defined moment in time (say a second), one person can know all that there is to know about the universe. since the universe is constantly changing, the future state cannot be stated definitely and one person cannot know what comes next in the future.

    In answer to my original question, I would assume no. My logic (which I feel has error) rests on the fact that in order to know everything, you must know with certainty that you know everything you don't know in order to make the comparison of how much you know. this is a paradox, therefore it is impossible.
  2. jcsd
  3. Oct 23, 2005 #2
    In terms of bits of information, we should be able to determine mathematically the total bits of information predicted in the universe from all possible interactions of the initial matter at time = 0--and recall, although the universe does change over time, if the big bang hypothesis holds, the universal total bits of information must remain a constant, with the bits only changing between matter and energy over time. Now, perhaps we also have models from neural biology that predict the total bits of information that a single human brain on average can hold. Then if the latter can hold the former, then the answer to your question would be, yes, a human could in theory have complete knowledge of everything in the universe, even that which it does not at this time know, but at some future time may come to know--, if the answer is no, then not. Here I use a definition of knowledge that is constrained to the facts and laws of matter plus energy = existence, as the basis of existence that can be known, and that the only way to "know" anything is by application of science, which by definition = "to know" (Webster). As I have stated before on another thread, if 100 % of what you hold to be true is from belief (e.g., not via science), then, by definition, you have 0.0% knowledge of that which in reality is true. Here I equate mathematical knowledge as a subset of scientific knowledge, but I would be open to other views.
  4. Oct 23, 2005 #3
    If one wishes to know the position of the cue ball on a billiard table, one needs to specify this with infinite precision. Anythng less than infinite precision will eventually result in error when trying to predict possible trajectories (chaos theory). Given that each decimal place requires finite time and space to specify, to specify any position with infinite precision is not possible in a universe which is finite in either space or time.

    So far we have only tried to specify the position of the cueball and found this impossible - and now we also need to know "everything"?

  5. Oct 23, 2005 #4


    User Avatar
    Science Advisor
    Gold Member

    Sounds like a similar question as to whether a "true" theory of everything is possible. The 'true' meaning that you obtain the knowledge needed to solve all scientific, philosophic and religious questions. I'd believe we can approach such answers by increasing our knowledge, but our best answers always only saturate near/towards the 'knowledge of everything'.
  6. Oct 23, 2005 #5
    Absolute and objective truth can never be achieved.
    In the final analysis, we are always a part of the universe we are trying to measure. This means that our perspective will always colour the "truth" that we perceive.
    Put into scientific terms : There is an epistemic horizon, characterised by the Heisenberg Uncertainty Principle, beyond which we can never see. Ultimate objective truth can always be pursued, but never reached.

  7. Oct 24, 2005 #6
    Is it possible to know everything?
    The same mind that is 'home to' ('creates' via concept, perception, memory, etc...) that which is being studied, is also formulating new things to study (as it has formulated the previous items for study); new problems to solve, as that is what the mind does; creates 'problems' (the 'material' universe)and 'solves' them (tries to understand it).

    The only universe you can 'know' is the one within your mind, and you can never know if there exists any actual 'in itself' thing in front of your nose! I see it working as an ever-spiral; imagining new 'particles' to discover, and then, lo and behold, discovering them; as an ever expanding dream.. Hence the 'consciousness' of the experimenter being integral within the experiment.

    So, 'Knowing' anything, in an absolute sense is not even possible, but knowing everything?
    One cannot ultimately know everything if one cannot ultimately, absolutely 'know' anything!
    All else is vanity...
  8. Oct 26, 2005 #7


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    You may notice that this thread has been pruned significantly. Please, let's try to keep the discussion at a high level of philosophical merit.
  9. Oct 28, 2005 #8


    User Avatar

    The only knowing everything
    is understanding all at once

    that the everpresent impermanence(creation)
    is all there is.

    Don't throw books at me.

  10. Nov 5, 2005 #9
    ? No philosophy defines "knowing" as "understanding all at once". And by what logic do you conclude that "creation" is impermanent ? It may be true, but logically it also may not be true, e.g., creation may last for infinity.
  11. Nov 5, 2005 #10


    User Avatar
    Gold Member

    This highlights an important distinction between Newton's deterministic universe and Heisenberg's "Uncertain" universe.

    In Newton's universe it is, at least in principle, possible to know the position and direction of every atom (and photon) to be arbitrary degree of accuracy. Knowing this means you also know every moment in the past and every moment in the future (a la a billiards shot writ large).

    HUP asserts that, even in principle, it is impossible to know everything.
  12. Nov 6, 2005 #11
    Except, of course, that to accomplish this we would need to know the position and direction of every atom (and photon) not only to an arbitrary degree of accuarcy, but to an infinite degree of accuracy. It would be impossible (even in principle) to encode this information within the universe, thus it would be impossible (in principle) to "know every moment in the past and every moment in the future", even without resorting to HUP. Chaos alone (in the absence of HUP) is enough to make the entire Newtonian spacetime "unknowable in principle"

  13. Nov 6, 2005 #12


    User Avatar
    Gold Member

    I would say 'impossible in practice, though not in principle'. I know of no rule that says explicitly that it can't be encoded. (perhaps with some really good lossless compression).

    On the other hand, HUP says explicitly, and in principle, that this level of knowledge cannot be achieved.
  14. Nov 6, 2005 #13
    Assuming Newtonian physics :
    To encode the position and velocity of one electron at one point in time, you would need at least 6 real numbers. Not 6 integers - 6 real numbers (one each for X, Y, Z coordinates, one each for dX, dY, dZ components of velocity). This probably would not be enough information, but let's stick with these 6 real numbers for the electron. The same needs to be done for every other electron and every other particle in the universe. Where are you going to store that information? If there are (sheer guess here) 10^120 particles in the universe you need to store 6 x 10^120 real numbers using those same particles...... it's impossible in principle (there is not enough storage space).

    Compression can only be applied where there is redundancy in information. We have no reason to believe that there is any redundancy in any of the 6 real numbers required for each particle - or do you know otherwise?

    Even if you could reduce all of the information required to just one real number, it still couldn't be stored. There is not enough storage space in a finite universe to store a complete real number. Unless you want to propose that all the information needed for the entire universe could be reduced to a single integer?

    Last edited: Nov 6, 2005
  15. Nov 6, 2005 #14


    User Avatar

    Look at your post.

    Do you see how logic(thought)
    about the non-existant is
    all false?
  16. Nov 6, 2005 #15


    User Avatar

    For example, you use the
    word "conclude"...that word implies permanence.

    Permanence does not exist.

    All philosophy is nonsense.
  17. Nov 6, 2005 #16


    User Avatar
    Gold Member

    I question your logic about compressibility.

    One example of compression where there is no redundancy:
    Two particles in space require those six coordinates to X decimal places of precision. That does not mean you need six numbers with X decimal places of precision. Recording merely the DELTA of the second particle (wrt the first) requires a way smaller number.

    Whether or not that's sufficient compression to overcime storage restrictions isn't the point. The point is, I've performed an arbitrary amount of compression without losing any data.

    Now, the overarching point here is this:

    We could discuss back and forth whether it's possible or not possible to record the info, but no matter which one of us appears to have the last word on the subject, it is always possible that there is something more to add ("Eureka, I came up with a better compression algorithm!") after a week, or a century.

    But there is no explicit, in principle, reason why it can't be done.

    Whereas HUP says there is.
  18. Nov 6, 2005 #17
    In fact this IS an example exploiting redundancy - if indeed the absolute positions of the particles is redundant information (as your example assumes). We can only dispense with the absolute positions and work just with deltas if the absolute position information is redundant.

    With 10^120 particles, talking deltas simply reduces the number of real numbers to 6 x ((10^120) - 1). Not much of a saving!

    In fact no saving at all when you remember that we are dealing with infinite precision reals - and (infinity - X) = (infinity)

    You have not compressed anything. (infinity - X) = (infinity)
    You still need an infinite storage space to store a real number - even if you could reduce all 6 x 10^120 reals to one real. Unless you can compress the real number to an integer? Are you claiming this?

    Now the overarching point here is: There is in principle no way to compress an infinite real number to an integer. And that will never change. Period.

  19. Nov 7, 2005 #18


    User Avatar
    Gold Member

    We are not dealing with infinities - that is a straw man. Restate your case without them. (BTW, no one - except you - said anything about integers either.)

    It is not necessary to record values to an infinite level of precision. It is only necessary to record them to an arbitrary level of precision. 'Arbitrary', in this case, being defined as "high enough to enable extrapolation of all future and all past trajectories and interactions".

    Again, impossible in practice, but not in principle.

    And again, as remains the case, debatable.

    Whereas, HUP (granting that it holds true), is not debatable, as it states the principle explicitly.
  20. Nov 7, 2005 #19
    No straw man. Try to specify an arbitrary real number EXACTLY by using integers. It cannot be done - in principle (not just in practice).

    Are you suggesting that you would use real numbers, and not integers, to store the data?

    "all future" would be an infinite level of precision, not simply an arbitrary level of precision. Use any combination of integers you wish to store this and eventually the prediction will be in error.

    Would you care to say "what" level of integer precision you would need?
    No, I guessed not. Because whatever level of precision you choose, it will result (eventually) in an inaccurate prediction (if we try to predict far enough ahead). The only level of precision that will "always" give an accurate prediction is an infinite level of precision.

    That HUP places a limit to our epistemic abilities is not in question - no need to keep repeating yourself here

  21. Nov 7, 2005 #20


    User Avatar
    Gold Member

    You are too quick on the draw with your comebacks. You are falling farther and farther off the mark. Nobody is saying "infinite", nobody is saying "exactly", and nobody is saying "integer".

    Or let me rephrase: you do not need to specify a coordinate exactly, or with infinite precision, nor using an integer. You need merely specify it to a level of precision that is good enough to get the job done.

    Until you're back on track, there's little point in addressing the rest of your comments.
  22. Nov 7, 2005 #21
    With respect, you are trying to wriggle off the hook by diluting the argument and avoiding my questions.

    Exactly what is "the job", in your opinion?

    Until you specify what you think "the job" is then there's little point in continuing the discussion

    Last edited: Nov 7, 2005
  23. Nov 7, 2005 #22


    User Avatar
    Gold Member

    If you've forgotten what we're talking about, why are you asking me? It's all there in black and white!

    You want me to review for you? Fine.

    The "job" is to know the position and trajectory of every particle at a given moment in time. If we live in a Newtonian universe, this is enough to know every event that precedes the present moment and every event that follows it as long as we know it to a high enough degree of accuracy. Thus, it would be knowing everything (is this sounding familiar?).

    Now, we don't have to know about a particle's postion and trajectory to an *infiinite* degree of accuracy (which you are claiming, not me), and we don't have to know it *exactly* (your claim, not mine), and we don't have to reduce it to integers (your claim, not mine). We need to know about it only to a degree of accuracy that is (granted, very, very high) high enough to confidently know where that particle will be a billion or trillion years (or whatever you want) in the future.

    OK, back on track?

    Now. Sure it's a huge undertaking, but there is no principle that you can can state that says "it cannot be done". There is no principle that says a system of X particles cannot EVER be described in less than Y bits of data.

    The question of whether the universe is big enough to record that information is certainly a valid one - and don't get me wrong - I grant that it is likely true. But it would be true in PRACTICE, not in principle.

    And here's the big thing:

    If I asked you to prove your principle holds, you could not do it. You could not prove your claim in principle, you could only prove it by actually doing it. By actually starting the measuring process.

    My whole point is that HUP shows in principle that you don't HAVE to demonstrate it in practice.

    Your argument amy be right, it is merely weak and nondemonstrable, based on opinion.

    And hey, how about a little space here? I'm meeting you more than half way in trying to keep this discussion on track. No more claming you need to stop for a review and pretending it's my fault eh?
    Last edited: Nov 7, 2005
  24. Nov 7, 2005 #23
    With respect, this shows your error in interpreting “what the job is” in this thread, and this is why I asked YOU to state what you THINK the job is, because your idea of what the job is clearly does not agree with mine. Read the first post in this thread once again. It quite clearly says :

    The “job” is thus NOT to know things “to a high enough degree of accuracy” – it is to know everything – ie to an UNLIMITED degree of accuracy. And the job is not simply “to know everything in principle” but to know it in practice. (the question did not specify “is it possible to know everything in principle?” – it simply specified “is it possible to know everything?”).

    Thus, we must ask : Are there any insurmountable impediments to our knowledge which would make it impossible for us to know everything? This means “practically insurmountable” impediments as well as “in principle insurmountable” impediments.

    The question is what is “high enough” to do the “job”? The “job” remember is “to know everything”, not simply to know most things approximately, or to an arbitrary degree of accuracy. If you simply wish to make approximate or arbitrary measurements and predictions, I can equally well argue that the HUP does NOT prevent me from doing the job – I can make approximate measurements and predictions even taking HUP into account. In fact, in the macroscopic world chaos is a much greater limitation than the HUP to accurate predictions. When you play billiards (pool), your inability to predict the ball’s trajectory more than a few collisions in advance is due to chaos, and NOT due to HUP. We rarely if ever encounter HUP limitations to our epistemology in the macroscopic world, but we almost always encounter chaos limitations to our epistemology.

    But making approximate measurements, I repeat, is not “knowing everything”.

    I hope we are now back on track. The “job” is to know everything, “all that there is to know”, and not simply to make an approximation to it.

    Now let’s look at specifics. How accurately do you wish to measure the position of each particle in our snapshot of the universe? In other words, to how many decimal places do you wish to measure the position? You obviously (refer to your above statement) do not want to measure to an infinite number of decimal places. Let’s say that you wish to measure it to P decimal places, where P is an integer. You may say that P can be arbitrarily large – I will show below that P must necessarily be very small.

    Now because you have “truncated” the real (true) position measurement by taking an arbitrary P number of decimal places in your measurement, this necessarily means that your future predictions will also be limited in their accuracy. Let us say that taking P decimal places in your original measurement of each particle allows you to predict those particle’s future positions at time t to an accuracy of Q decimal places, where Q is another integer. There will be a 1:1 mapping between values of P and Q – for any given integer P decimal places of original measurement there will be a limit to the accuracy of your future prediction at time t of Q decimal places. Yes P can be “arbitrarily large” if you have enough storage espace, but P is always an integer and Q is always an integer. There is an a priori limit to the precision of your predictions imposed by the value of P that you choose. This is an “in principle” limit, not simply a practical limit, imposed by your choice of P.

    The same argument applies to measuring the velocity of the particle. The “real” velocity is a real number, and you must necessarily truncate this to P decimal places when you measure and record that number. The P decimal place number that you store is an approximation to the real velocity of the particle.

    “So what” I can hear you say – “I can make P as large as I like – I can make it arbitrarily large”.

    But can you? You now need to record these P decimal places of position for EVERY particle in the universe (the job is to know everything, remember?). If there are N particles in the universe this means you will need to record 6xN numbers (one each for the X, Y, Z coordinates, one each for the dX, dY, dZ components of velocity), and each of these numbers to to P decimal places of accuracy. Thus (assuming you are working in base 10) you will have 6xNxP base 10 digits to store and to work with. N is a VERY large number – possibly as large as 10^87.

    Now you say “But I can compress the data by working with relative positions and velocities rather than absolute positions and velocities”

    This saves you just 6xP digits out of your total of 6xNxP digits. So now you need to store and work with just 6x(N-1)xP digits.

    Now the big question, which gets us back to how large P can possibly be : Where are you going to store all this information? You only have the universe itself to use as your storage device after all. Thus your “job” is to store 6x(N-1)xP decimal digits using just N particles. If P is equal to 1 or more decimal places, you simply will not enough enough storage space – not just in practice but in principle. Once again this is an “in principle” limit – not simply a practical limit. Let’s put some numbers in to “see” the problem more clearly.

    How accurately would you like to measure the positions and velocities? To 5 decimal places perhaps? This seems reasonable (but it would be grossly insufficient to then allow us to claim that “we know everything” – because in fact we would know and predict to only 5 decimal places or less, we would plainly not know everything).

    OK, so let’s assume P is 5. N is probably around 10^87. Thus we end up having to store 6 x (10^87 -1) x 5 decimal digits, which is effectively equal to 3 x 10^88 digits. We must store these digits using 10^87 particles (I am not even going to bother going into the practical problems of storing just 1 digit using 1 particle). It simply cannot be done – it is impossible in principle. Using 10^87 particles and ignoring practical storage problems, we could not encode more than 10^87 digits (in fact the number would be much less than 10^87 digits) – but this means that our storage limitation – a limitation in principle not just in practice – forces us to record each position and velocity to LESS THAN one decimal place.

    The only possible solution you might have at your disposal is to be able to store more data IF the data can be compressed – but this reqires an incredible amount of redundancy in the data, and we have NO reason to believe there is any redundancy in the relative position or velocity data of the particles involved. But no matter how much you compress the data, you will always necessarily store an integer number of decimal places for each particle - whereas the real particle's postion and velocity are denoted by real numbers each with an infinite number of decimal places. No matter how much you compress the data, infinity always remains infinity. You simply cannot, in principle, store the "real" position data, the best you can ever do is to store an "approximation to it". Thus there is ALWAYS an in principle impediment to "knowing everything".

    I will cut some slack here and I will agree that the absence of data redundancy is a practical limit at the present time, rather than an in principle limit. But there is no reason to believe, given what we know about the universe, that this limit will ever be overcome. And even if it can be overcome, the real data is a real number and can NEVER be compressed to an integer number of decimal places. And this, dealing as it does with the "definitional differences" between real numbers and integers, is an In Principle limit and not merely a practical one.

    We can continue to argue whether the lack of storage is an in principle or a practical limitation – but I will also remind you that the original question did not ask whether it is possible to know everything “in principle” – it simply asked whether it is possible.

    Thus the correct answer to the question posed at the beginning of this thread, which was “Is it possible to know everything?”, is “No, for two reasons – lack of storage space, and the HUP”.

    Last edited: Nov 8, 2005
  25. Nov 8, 2005 #24


    User Avatar
    Gold Member

    Yup. Totally agree.
    Lack of storage space would be a practical limit, while HUP limits it in principle.

    HUP is a strong argument because it can be shown - before any calculations are even done - that the attempt is futile. Storage space is a weaker argument (though valid) in that one cannot inarguably state that "you cannot record X data in Y space" or somesuch.
  26. Nov 9, 2005 #25

    I can't believe you actually got responses. The answer is No. You can't even know everything about 1 atom. How are you gonna know everything about the rest of the U?
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook