1. Not finding help here? Sign up for a free 30min tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Apparently I just dont understand this

  1. Nov 20, 2007 #1
    I am extremely confused right now, concerning the limits and I guess the limit rules... I'm either reading it wrong or its just not explaining it well enough for me to understand. Basically, I am lost on this entire chapter of limits. This is kinda what I think is going on.

    There are limits to finite series and a finite series is a series of numbers that continue up until a stopping point or number, like 1,2,3,4...10. Then there are infinite series, which never end, but converge to a number, which I guess would be infinite, so in my mind, these partial sums, which are always getting smaller(?), because they are fractions(?), will never reach this point, and become an infinitesimal, which is a number so small that Thompson says it can be 'thrown out'.

    However, it says in this book that these numbers can go beyond this 'limit', somehow, in an infinite series. My first question is how is this possible?

    My understanding of this whole limit of infinite series is (and is probably wrong, because I'm totally confused), is that lets say I have a square. Corner A of this square to corner B of this square is 10 yards. This 10 yards is the limit, of an infinite number of squares in a row, perhaps. So first I go, 5 yards, then 2.5 yards, etc. I will never reach this particular SQUARE'S limit..but it doesn't or isn't relevant to the infinite number of squares because we are solving whatever it is we're trying to solve for this one particular 'limit' square.

    So is that at least partly right?

    Lastly, the book gave me some examples of what the Gardner calls, "the integral limit of any repeating decimal."

    So my last question is what is the point of solving this integral limit of any decimal?
  2. jcsd
  3. Nov 20, 2007 #2


    User Avatar
    Staff Emeritus
    Science Advisor

    Are you clear on the distinction between "sequences" and "series"? A sequence is a list of numbers: 1, 2, 3, 4, ... 10 is a finite sequence of numbers. If an infinite sequence "converges to a number" it would not be infinite. "converging" always means "getting close and closer to a finite number:. The infinite list of numbers, 1/2, 2/3, 3/4, 4/5, ..., n/(n+1) converges to 1. Notice that the numerator is always less than the denominator so every number in the sequence is less than 1. If you divide both numerator and denominator by n, you get 1/(1+ 1/n). As n gets larger and larger that fraction 1/n gets smaller and smaller (the limit of the sequence 1, 1/2, 1/3, ..., 1/n is 0).
    But then you talk about "partial sums". A "series" is a sum of numbers. You can always add a finite sequence, but most infinite sequences cannot be added- the "series" does not converge. The "partial sums" of an infinite series are the finite sums. for the series
    [tex]\Sum_{n=1}^\infty \frac{n}{n+1}[/tex]
    The partial sums would be 1/2, 1/2+ 2/3= 7/6, 1/2+ 2/3+ 3/4= 23/12, etc. Those numbers are clearly getting larger and larger and the series does NOT "converge". It is possible to show that if a sequence does not converge to 0 then the corresponding series (sum) does not converge.
    [tex]\Sum_{n=1}^\infty \frac{1}{n}[/itex]
    has the property that, even though the individual terms go to 0, the series does not converge.
    [tex]\Sum_{n=1}^\infty \frac{1}{2^n}[/itex]
    on the other hand, does converge. It converges to 1.

    I must say that I dislike talking about "infinitesmals" in a stituation like this. It is possible to do it rigorously but that requires some deep results from logic.

    I'm afraid I don't know how to answer that because I don't know what 'limit' you are talking about or what you mean by going "beyond" a limit.

    You are confusing things slightly by talking about "squares". You really are talking about a line segment. Yes, if you go 5 yards along the 10 yard line segment, then 2.5 yards, then 1.25, etc., after any finite number of steps you have not gone the entire distance but in the limit you will reach a total of 10 yards. That is, in fact, precisely the example I gave before, multiplied by 10:
    [tex]\Sum_{n=1}^\infty \frac{10}{2^n}= 10(1)= 10[/itex]
    Both this example and mine are "geometric series".

    I can't make heads or tails out of "but it doesn't or isn't relevant to the infinite number of squares because we are solving whatever it is we're trying to solve for this one particular 'limit' square.". What do you mean by "solving whatever it is we're trying to solve"?

    A little bit to unclear to be either right or wrong!

    "the integral limit of any repeating decimal." Could you post some of these examples? The only limit I can connect with a repeating decimal is the infinite series given by its digits: .333...= .3+ .03+ .003+ .0003+ ... That's also a "geometric series and it's easy to show that that converges to 1/3. Indeed, the definition of "decimal representation of a number" is that it is the sum of the infinite series formed by its digits.
  4. Nov 20, 2007 #3
    Hmm...well like I said I dont really know what Im talking about haha. Well, sorry for the misunderstanding, I got sequences and series confused. This is what I thought a series was, I thought a series was a particular 'set' of sequences. For example I thought a series could be 2,3,4,5, and 6 out of 1-10. But apparently not. So thats cleared up.

    The 'going beyond the limit' thing is what I didnt get at all, heres a direct quote from the book,

    "The partial sums go beyond the limit before they converge".

    How does that happen? I mean, perhaps earlier I didnt make it clear enough for you to know what I meant, so I am sorry about that.

    What I meant by 'solving whatever it is we're trying to solve', is what exactly is the answer we're looking for here? Im not sure how to make this clear to you, since obviously I dont know what Im doing, but are there some sort of problems like this in calculus? I mean, involving limits? I cant really give you an example, because I dont know of any...so sorry about the misunderstanding there as well.

    And for the integral limit, Gardner puts "There is a ridiculously easy way to determine the integral limit of any repeating decimal." Before that he uses "Limits of infinite series". And yes, he uses the example you gave, .3333 or 3/10, 3/100, 3/1000...=1/3.

    I hope that is clearer for you, I really didnt know how to put it, haha. Sorry about that..
  5. Nov 18, 2009 #4
    I believe that what the book is saying about reaching the limit before converging to it involves, for example, a series whose partial sums are 0, 1, 2, 1.5, 1.25, 1.125, 1.0625, 1.03125, etc. In this case, the limit of the series, 1, is also one of its partial sums.
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?

Similar Discussions: Apparently I just dont understand this