Could someone provide me with a good explanation of why $0^0=1$?

My train of thought:

$x>0$

$0^x=0^{x-0}=0^x/0^0$, so

$0^0=0^x/0^x=\,?$

Possible answers:

  1. $0^0\cdot0^x=1\cdot0^x$, so $0^0=1$
  2. $0^0=0^x/0^x=0/0$, which is undefined

PS. I've read the explanation on mathforum.org, but it isn't clear to me.

  • @Nuno: I agree in general, and in this case. I've edited the tags; Stas said in his comments on my answer that he meant the question in "terms of discrete mathematics", so the discrete-mathematics tag is appropriate, and (powers), (zeros), and (number-theory) are all inappropriate. – Arturo Magidin Nov 21 '10 at 3:15
  • 1
    Can you pls link to the explanation that you read on mathforum.org? – Lazer Nov 21 '10 at 6:48
  • 2
    BTW, this is comprehensively covered on Wikipedia (or see today's version), along with pointers to history and treatment in many systems. – ShreevatsaR Nov 22 '10 at 12:34
  • see also math.stackexchange.com/questions/135/… – sdcvvc Feb 11 '13 at 21:59
  • 4
    @Stas : actually, $0^0$ is generally considered "undefined". But when people use power series, they routinely treat $0^0$ as $1$ without a second thought. If a function $f$ is defined by a power series as $f(x) = \sum_{n=0}^\infty a_n (x-b)^n$, then everyone agrees that $f(b) = a_0$, even though plugging in $x = b$ into the series involves $0^0$. I hope this adds something to the dozens of other comments and previous answers. – Stefan Smith Apr 15 '13 at 16:26

26 Answers 26

up vote 309 down vote accepted

In general, there is no good answer as to what $0^0$ "should" be, so it is usually left undefined.

Basically, if you consider $x^y$ as a function of two variables, then there is no limit as $(x,y)\to(0,0)$ (with $x\geq 0$): if you approach along the line $y=0$, then you get $\lim\limits_{x\to 0^+} x^0 = \lim\limits_{x\to 0^+} 1 = 1$; so perhaps we should define $0^0=1$? Well, the problem is that if you approach along the line $x=0$, then you get $\lim\limits_{y\to 0^+}0^y = \lim\limits_{y\to 0^+} 0 = 0$. So should we define it $0^0=0$?

Well, if you approach along other curves, you'll get other answers. Since $x^y = e^{y\ln(x)}$, if you approach along the curve $y=\frac{1}{\ln(x)}$, then you'll get a limit of $e$; if you approach along the curve $y=\frac{\ln(7)}{\ln(x)}$, then you get a limit of $7$. And so on. There is just no good answer from the analytic point of view. So, for calculus and algebra, we just don't want to give it any value, we just declare it undefined.

However, from a set-theory point of view, there actually is one and only one sensible answer to what $0^0$ should be! In set theory, $A^B$ is the set of all functions from $B$ to $A$; and when $A$ and $B$ denote "size" (cardinalities), then the "$A^B$" is defined to be the size of the set of all functions from $A$ to $B$. In this context, $0$ is the empty set, so $0^0$ is the collection of all functions from the empty set to the empty set. And, as it turns out, there is one (and only one) function from the empty set to the empty set: the empty function. So the set $0^0$ has one and only one element, and therefore we must define $0^0$ as $1$. So if we are talking about cardinal exponentiation, then the only possible definition is $0^0=1$, and we define it that way, period.

Added 2: the same holds in Discrete Mathematics, when we are mostly interested in "counting" things. In Discrete Mathematics, $n^m$ represents the number of ways in which you can make $m$ selections out of $n$ possibilities, when repetitions are allowed and the order matters. (This is really the same thing as "maps from $\{1,2,\ldots,m\}$ to $\\{1,2,\ldots,n\\}$" when interpreted appropriately, so it is again the same thing as in set theory).

So what should $0^0$ be? It should be the number of ways in which you can make no selections when you have no things to choose from. Well, there is exactly one way of doing that: just sit and do nothing! So we make $0^0$ equal to $1$, because that is the correct number of ways in which we can do the thing that $0^0$ represents. (This, as opposed to $0^1$, say, where you are required to make $1$ choice with nothing to choose from; in that case, you cannot do it, so the answer is that $0^1=0$).

Your "train of thoughts" don't really work: If $x\neq 0$, then $0^x$ means "the number of ways to make $x$ choices from $0$ possibilities". This number is $0$. So for any number $k$, you have $k\cdot 0^x = 0 = 0^x$, hence you cannot say that the equation $0^0\cdot 0^x = 0^x$ suggests that $0^0$ "should" be $1$. The second argument also doesn't work because you cannot divide by $0$, which is what you get with $0^x$ when $x\neq 0$. So it really comes down to what you want $a^b$ to mean, and in discrete mathematics, when $a$ and $b$ are nonnegative integers, it's a count: it's the number of distinct ways in which you can do a certain thing (described above), and that leads necessarily to the definition that makes $0^0$ equal to $1$: because $1$ is the number of ways of making no selections from no choices.

Coda. In the end, it is a matter of definition and utility. In Calculus and algebra, there is no reasonable definition (the closest you can come up with is trying to justify it via the binomial theorem or via power series, which I personally think is a bit weak), and it is far more useful to leave it undefined or indeterminate, since otherwise it would lead to all sorts of exceptions when dealing with the limit laws. In set theory, in discrete mathematics, etc., the definition $0^0=1$ is both useful and natural, so we define it that way in that context. For other contexts (such as the one mentioned in mathforum, when you are dealing exclusively with analytic functions where the problems with limits do not arise) there may be both natural and useful definitions.

We basically define it (or fail to define it) in whichever way it is most useful and natural to do so for the context in question. For Discrete Mathematics, there is no question what that "useful and natural" way should be, so we define it that way.

  • 8
    @Sivam: exactly as Qiaochu says. Note that I said that $0^0=1$ in cardinal exponentiation is the only sensible answer, but "cardinal exponentiation" is not the same as real number exponentiation; when doing real number exponentiation, $0^0$ is most properly undefined/indeterminate. – Arturo Magidin Nov 20 '10 at 23:30
  • 14
    Just a small note: the answer depends on whether you think of exponentiation as a discrete operation (as in set theory, algebra, combinatorics, number theory) or as a continuous operation over spaces like real/complex numbers (as in analysis). – Kaveh Nov 21 '10 at 5:47
  • 22
    "In Calculus and algebra, there is no reasonable definition", should you really include algebra there? Aren't polynomials considered part of algebra, and don't people say that evaluating the polynomial $\sum_{k=0}^n a_k x^k$ at $0$ gives $a_0$? – Omar Antolín-Camarena Jun 8 '13 at 20:28
  • 22
    It is not usually left undefined. It is defined as $1$. We write $f(x)=\sum_{i=0}^\infty a_ix^i$ for a power series, but know that $f(0)=a_0$. In set theory, the definition is simple. In lambda calculus likewise. There is only one value of $0^0$ that makes sense, and it is $1$, and we use this quite a lot when we write polynomials and power series. While it is defined, it is still an "indeterminate form." Being indeterminate is a question of continuity. – Thomas Andrews Nov 1 '14 at 15:23
  • 22
    I wish I could downvote this answer several times. Saying that in algebra there's no reasonable definition is absurd. – egreg Nov 19 '14 at 11:49

This is merely a definition, and can't be proved via standard algebra. However, two examples of places where it is convenient to assume this:

1) The binomial formula: $(x+y)^n=\sum_{k=0}^n {n\choose k}x^ky^{n-k}$. When you set $y=0$ (or $x=0$) you'll get a term of $0^0$ in the sum, which should be equal to 1 for the formula to work.

2) If $A,B$ are finite sets, then the set of all functions from $B$ to $A$, denoted $A^B$, is of cardinality $|A|^{|B|}$. When both $A$ and $B$ are the empty sets, there is still one function from $B$ to $A$, namely the empty function (a function is a collection of pairs satisfying some conditions; an empty collection is a legal function if the domain $B$ is empty).

  • 45
    You don't need to appeal to the binomial formula. Anytime you write a polynomial as f(x) = sum a_i x^i you need x^0 = 1 to keep your notation consistent, so you need 0^0 = 1 so that f(0) = a_0. – Qiaochu Yuan Nov 21 '10 at 1:12
  • 9
    Yes. I think it is reasonable to define $0^0=1$ (because that seems to be the most useful definition) with the caveat that the function $x^y$ on $\mathbb{R}^{+}\!\!\times\mathbb{R}$ is not continuous at $(0,0)$. – robjohn Nov 13 '13 at 11:28

$0^{0}$ is just one instance of an empty product, which means it is the multiplicative identity 1.

I'm surprised that no one has mentioned the IEEE standard for $0^0$. Many computer programs will give $0^0=1$ because of this. This isn't a mathematical answer per se, but it's worth pointing out because of the increasingly computational nature of modern mathematics, so that one doesn't run afoul of anything.

The use of positive integer exponents appears in arithmetic as a shorthand notation for repeated multiplication. The notation is then extended in algebra to the case of zero exponent. The justification for such an extension is algebraic. Furthermore, in abstract algebra, if $G$ is a multiplicative monoid with identity $e$, and $x$ is an element of $G$, then $x^0$ is defined to be $e$. Now, the set of real numbers with multiplication is precisely such a monoid with $e=1$. Therefore, in the most abstract algebraic setting, $0^0=1$.

Continuity of $x^y$ is irrelevant. While there are theorems that state that if $x_n \to x$ and $y_n \to y,$ then $(x_n + y_n) \to x+y$ and $(x_n)(y_n) \to xy$, there is no corresponding theorem that states that $(x_n)^{(y_n)} \to x^y$. I don't know why people keep beating this straw man to conclude that $0^0$ can't or shouldn't be defined.

  • 6
    Downvote, because Onez focuses on a very narrow view of the exponentiation operation and its applications, and writes as if that is the only view. Continuity is of rather significant importance in a wide variety of situations, and the requirements of an exponentiation function of continuous arguments are rather different than those limited to integer or rational exponents, and $0^0$ runs into that difference. Another example where the needs differ are $(-1)^{1/3}$ – Hurkyl Oct 26 '11 at 22:06
  • 8
    I hardly consider the whole domain of algebra to be narrow. YMMV. In any case, when extending the domain of functions, one may ask: Is the extension useful? Defining 0^0 to be 1 is useful in Combinatorics, Set Theory, and Algebra. Indeed, it is even useful in calculus when using summation notation for polynomial functions and infinite series. Perhaps you care to list several advantages of leaving 0^0 undefined? Particularly in light of the fact that many definitions require the additional caveat that a,b,x,y etc. not be equal to zero in order for them to be true. – Onez Oct 27 '11 at 1:13
  • 7
    By your argument, we should not define (-2)^0 = 1, but leave it undefined since this extension of exponentiation fails your second case (non-positive base) and your third case (no continuous extension of x^y to all of C). You still haven't supplied any advantages to leaving 0^0 undefined. Economy of notation (the reason exponential notation was developed in the first place) is gained by defining 0^0 = 1. – Onez Oct 27 '11 at 21:11
  • 5
    Since 0 is not in the range of the exponential function, I take it you have issue with 0^y being defined even for y>0. The argument seems to hinge on whether one is to define 0^0=1 and economize several definitions and theorems from algebra, combinatorics, and analysis, at the expense of one caveat for a single function, OR to leave 0^0 undefined, have several caveats so as to preserve the continuity on the domain of definition of a single function, namely x^y. Where is the greatest economization to be had? Who has the narrow view? – Onez Oct 28 '11 at 3:29
  • 7
    You missed the more relevant option -- acknowledge the multiple exponentiation operations that come up in mathematics, rather than conflating them. – Hurkyl Oct 28 '11 at 6:18

Maybe it's a good idea to put the problem into a broader perspective.

The minimum we need to define a power is a multiplicative semigroup, which basically means we have a set with an associative operation which we write as multiplication. If $S$ is the semigroup, the power function is defined as $$S\times \mathbb Z^+ \to S, (x,n) \mapsto x^n = \begin{cases} x & n=1\\ x x^{n-1} & n>1 \end{cases}$$ This power function has the fundamental properties $$x^mx^n = x^{m+n},\quad (x^m)^n = x^{mn}\tag{*}$$ Note that up to now, we have used absolutely nothing about the elements of $S$ except for the fact that we can multiply them.

Now it is a natural question whether we can extend that definition from the positive integers to the non-negative integers, in other words, whether we can define $x^0$. Of course we would want to define it in a way that the power laws (*) still hold. This especially implies the following relations: $$(x^0)^2 = x^0 \tag{I}$$ (that is, $x^0$ is idempotent) and $$x^0 x = x x^0 = x.\tag{II}$$ Note that this does not imply that for any $x\ne y$, $x^0=y^0$.

Now for a given $S$ and a given $x\in S$, there are three possible cases:

  1. There does not exist an element $z\in S$ that fulfils both (I) and (II). In that case, $x^0$ is undefined and undefinable without violating the power laws. For example, if you take $S$ as the set of positive even numbers, then $x^0$ is undefined for all $x\in S$ (because there's no positive even number that fulfils either (I) or (II) for any x)
  2. There exists exactly one element $z\in S$ that fulfils those conditions. In that case, the only reasonable definition is $x^0 = z$. For example, for non-zero real numbers, you get $x^0=1$ this way.
  3. There exist more than one element $z_i\in S$ fulfilling those equations. This is the case for $0^0$ because both $0$ and $1$ fulfil the equations. In that case, you have several choices:

    • You can select one of the possible values and define $x^0$ as that value. Of course you'd not randomly choose any value, but choose the one which is most useful. Which may be different in different contexts. Note that each of the choices gives a different valid power function.
    • You can leave $x^0$ undefined. In this case, $x^0$ is called unspecified because you could specify it (as in the previous bullet). In particular, the restriction of any of the power functions from the first bullet to $S\setminus\{x\}$ will be the power function from this bullet.

Now one particularly interesting case is if you have a neutral element, that is, an element $e\in S$ so that $ex=xe=x$ for all $x\in S$. A semigroup with such an element is called a monoid.

It is easy to check that in this case, $e$ fulfils both (I) and (II) for any element $x\in S$. Therefore you always get a valid power function by defining $x^0=e$. Indeed, by doing so, you get the general form $$x^n = \begin{cases}e & n=0\\ x x^{n-1} & n>0\end{cases}$$ which means that in a monoid, $x^0=e$ clearly is a distinguished, and therefore preferable choice of the power function, even for $x$ where other choices would be possible. Note that the real numbers form a monoid under multiplication, with $e=1$. Therefore this is a first hint that $0^0=1$ is a good definition.

The definition $0^0=1$ turns out to be useful also in other areas, for example when considering the linear structure (that is, the distributive law with addition), as it makes sure that e.g. the binomial formula $$(a+b)^n = \sum_{k=0}^n \binom{n}{k} a^k b^{n-k}$$ also holds if any of $a$, $b$ or $a-b$ is $0$.

There's however one structure that doesn't favour $0^0=1$, and that is continuity: $x^y$ is discontinuous at $(0,0)$, and every value can be obtained as limit of a suitably chosen sequence of arguments approaching the origin. Therefore from the point of continuity, $0^0$ can be considered unspecified. However, since $x^y$ is discontinuous at the origin whether or not one defines it there, and independent of the value one chooses, that is not really an argument against choosing $0^0=1$, but more an argument against using that definition blindly.

I'm not aware of a context where the definition $0^0=0$ would be more useful, but I wouldn't want to exclude that it exists. Any other value of $0^0$ would violate condition (I) above, and therefore likely not be useful at all.

So in summary, the definition $0^0=1$ is the most useful in most situations, and not harmful in situations where one would otherwise leave $0^0$ unspecified, and if any context exists where another definition would be more useful, it's arather unusual context. Therefore the definition $0^0=1$ is the most reasonable one.

$0^0$ is undefined. It is an Indeterminate form.

You might want to look at this post.

Why is $1^{\infty}$ considered to be an indeterminate form

As you said, $0^0$ has many possible interpretations and hence it is an indeterminate form.

For instance,

$\displaystyle \lim_{x \rightarrow 0^{+}} x^{x} = 1$.

$\displaystyle \lim_{x \rightarrow 0^{+}} 0^{x} = 0$.

$\displaystyle \lim_{x \rightarrow 0^{-}} 0^{x} = $ not defined.

$\displaystyle \lim_{x \rightarrow 0} x^{0} = 1$.

  • 15
    Is $\lim_{x\to0}0^x$ really defined? It can only be approached from the positive side. – kennytm Nov 21 '10 at 19:52
  • 2
    @KennyTM: Accepted and edited accordingly. – user17762 Nov 22 '10 at 11:47
  • 5
    0^0 is undefined by whom? I saw somebody defined it, can I now say it is defined? – Anixx Nov 4 '12 at 22:58
  • 5
    This answer is wrong. It is indeed an indeterminate form, but it is defined and equal to $1$. See my answer also posted here. – Michael Hardy Mar 5 '14 at 18:06
  • 7
    Being an indeterminate form is not the same as being undefined. They are two different concepts. – Thomas Andrews Nov 1 '14 at 15:25

It's pretty straight forward to show that multiplying something by $x$ zero times leaves the number unchanged, regardless of the value of $x$, and thus $x^0$ is the identity element for all $x$, and thus equal to one.

For the same reason, the sum of any empty list is zero, and the product is one. This is when a product or sum of an empty list is applied to a number, it leaves it unchanged. Thus if the product $\Pi()$ = 1, then we immediately see why $0! = 0^0 = 1$.

Without this property, one could prove that $2=3$, by the ruse that there are zero zeros in the product on the left (zero is after all, a legitimate count), and thus $2*0^0$, and since $0^0$ as indeterminate, could be 1.5, and thus $2=3$. I think not.

The approaches to $0^0$ by looking at $x^y$ from different directions, fails to realise that for even lines close to $x=0$, the line sharply sweeps up to 1 as it approaches $y=0$, and that the case for $x=0$, it may just be a case of not seeing it sweep up. On the other hand, looking from the other side, even in a diagonal line (ie $(ax)^x$), all do rapidly rise to 1, as x approaches 0. It's only when one approaches it from $0^x$ that you can't see it rising. So the evidence from the graph of $x^y$ is that $0^0$ is definitely 1, except when approached from $y=0$, when it appears to be zero.

"Everybody knows" that $$ e^z = \sum_{n=0}^\infty \frac{z^n}{n!}, $$ and when $z=0$ then the first term is $\dfrac{0^0}{0!}$, so of course $0^0$ is $1$ since it's an empty product.

But it's also an indeterminate form because $\displaystyle\lim_{x\to z}f(x)^{g(x)}$ can be any positive number, or $0$ or $\infty$, depending on which functions $f$ and $g$ are, if $f(x)$ and $g(x)$ both approach $0$ as $x\to a$.

Knuth's answer is at least as good as any answer you're going to get here: http://arxiv.org/pdf/math/9205211v1.pdf See pp. 4-6, starting at the bottom of p. 4.

Another reason to define $0^0=1$ comes from probability, in particular Bernoulli trials.

These are independent repeated trials of an experiment, whose outcome is either positive or negative. So let $p$ be the probability of a success for each trial. Then the probability of exactly $k$ successes out of $n$ trials is $$p_k = {n \choose k}p^k(1-p)^{n-k}.\tag{B}$$ Now, suppose $p=1$. $(B)$ yields $p_n=0^0$. However, we already know that each trial will certainly occur, resulting in $p_k=0$ for all $k<n$ and $p_n=1$, whence $$0^0=1.$$

It depends on whether the 0 in the exponent is the real number 0, or the integer 0. These are two different objects and while the distinction is not often important, it is in this case.

Exponentiation by an integer has a universal specific meaning: Positive exponent is repeated multiplication, negative exponent is the inverse of repeated multiplication, zero exponent is the empty product, equal to the multiplicative identity denoted by 1. So if the exponent is the integer 0, then $0^0=1$ - the fact that this is an empty product with no terms trumps the fact that the terms which are not there are 0's (because they're not there).

Whereas exponentiation by a real or complex number is a messier concept, inspired by limits and continuity. So $0^0$ with a real 0 in the exponent is indeteriminate, because you get different results by taking the limit in different ways.

Note that all of the standard examples where it's "convenient" to have $0^0=1$ (e.g. power series) are all cases where the exponent is an integer.

Note that as long as the exponent is the integer 0, it doesn't matter if the base is the integer 0, the real 0, or pretty much any mathematical object on the planet for which multiplication is defined.

  • @Wildcard. Thanks. This is one of those things that unfortunately too few people understand. That's what happens when people limit themselves to thinking of real numbers and fail to realize that math is much richer than that - with all kinds of structures which look nothing like numbers, and yet where multiplication and integer exponentiation make sense. Myself I've picked up this distinction from Mike Oliver, aka User:Trovatore on Wikipedia, and I'm happy to spread the word. – Meni Rosenfeld Jul 12 '17 at 13:37
  • @Meni Useful distinction. But FWIW, problematic in my view. Saying that the integer exponent zero is somehow different than the real number exponent zero is tricky at best. How would you even write that symbolically? I bet expressions like $0 \in \mathbb{R}$ or $0 \in \mathbb{Z}$ would be hard to find in mathematical papers. As a software developer the concept of a typed zero makes sense to me, but I'm confident that it would be rare to find languages or run-time libraries where $0^{0.0} \ne 0^0$. – Χpẘ Jul 26 '17 at 17:18
  • @Χpẘ: You are correct that the standard notation does not make it easy to draw this distinction. However, this is more of a conceptual matter, and I believe that in any case where the distinction matters, the correct interpretation will be clear from context. As for libraries, I don't know which actually do that, but they totally should return NaN for $0^{0.0}$ and 1 for $0^0$. – Meni Rosenfeld Jul 26 '17 at 21:35
  • @Χpẘ zeros are always problematic, because they wipe out unit multipliers. So zero miles equals zero gallons equals zero light-year kilotons per microsecond equals a zero wavelength equals zero time and zero space...and now we're in the realm of philosophy. Zero is the most important number to be "typed." – Wildcard Aug 8 '17 at 9:27
  • @Wildcard Interesting. In my intuition zero siriometers equals zero cow's grasses equals zero shakes, etc. They all signify no-thing. At least in realms concerned with dimensional analysis. In usage outside of such realms, it would be strange to say, "We've got zero kinematic viscosity in the fuel tank - we're not going anywhere." But English has plenty of inconsistencies. That said, I don't see how this relates to a definition of $0^0$. You can rewrite dimensional equations to be dimensionless. Additionally Wikipedia says: Scalar args to transcendental functions must be dimensionless. – Χpẘ Aug 9 '17 at 22:26

There are so many fantastic answers already, but no pictures which is sad. $0^0$ is undefined - is it the limited case of $x^0$, $0^x$ or $x^x$? The plots for $x^0$ (a line at 1) and $0^x$ (a line at zero) are boring, but $x^x$ is

$x^x$

Author/Site: Wolfram|Alpha

Publisher: Wolfram Alpha LLC

URL: https://www.wolframalpha.com/input/?i=plot+x%5Ex

Retrieval date: 6/2/2015

I don't always define $0^0$, but when I do, I define it as $1$.

  • 3
    To say something is undefined you should be sure nobody defined it. – Anixx Feb 7 '16 at 11:24
  • @selfawareuser Did you just defend an idea in math by dividing by zero? Look at the implicit limit you took and you'll find that I covered that case already without dividing by zero. Don't divide by zero. – user121330 Aug 5 '16 at 20:25
  • @user121330 The prohibition is against dividing a finite number by zero. $\frac{0}{0}$ in itself is legal but you must say $\frac{0}{0}\rightarrow x$ i.e. substitution is unidirectional. – user301988 Aug 5 '16 at 21:35
  • @selfawareuser Lol, so since $\frac{0}{0} = \frac{1 \cdot 0}{3 \cdot 0}$, one might also claim that $\frac{0}{0} = \frac{1}{3}$? Why don't you provide a citation validating your claim. Better yet, start with $y^{x-x}$, and follow your algebra to find that you really just took the limit of $y^0$ as $y$ approaches zero which is still one. – user121330 Aug 8 '16 at 13:55
  • Citation, from my user history. – user301988 Aug 9 '16 at 7:41

Some indeterminates forms $0^{0}, \displaystyle\frac{0}{0}, 1^{\infty}, \infty − \infty, \displaystyle\frac{\infty}{\infty}, 0 × \infty, $ and $\infty^{0}$

Futhermore,

$$\lim_{ x \rightarrow 0+ }x^{0}=1$$ and

$$\lim_{ x \rightarrow 0+ }0^{x}=0$$

See http://en.wikipedia.org/wiki/Indeterminate_form

  • 16
    One should note that indeterminate is not the same as undefined. – Hagen von Eitzen Jul 10 '13 at 16:27
  • 10
    @JMCF125: $\lim\limits_{(x,y)\to(0,0)}x^y$ is indeterminate. This actually frees us to define $0^0$ to be whatever value is most useful. In almost every practical case, that is $0^0=1$. – robjohn Feb 10 '14 at 6:23
  • 3
    @JMCF125: That's why the site is here. Where is the related question? – robjohn Feb 10 '14 at 14:25
  • 2
    @robjohn, this one. – JMCF125 Feb 10 '14 at 19:37
  • 2
    @Hagen the limit of $0^x$ as x-> 0 is the same as 0/0, which is undefined, but $0^0$ can be evaluated without division by 0, and these proofs lead directly to $0^0=1$. – wendy.krieger Oct 13 '14 at 4:31

Another approach and yet another result !

We define an exponential function on $\mathbb{R}$ as a function $E:\mathbb{R}\rightarrow \mathbb{R}$ such that $$ E(x+y)=E(x)E(y) \qquad \forall x,y \in \mathbb{R} $$ That's a very general and powerful definition, and I love it because captures the link between Lie algebras and Lie groups.

From this definition we have immediately that if $\exists a\in \mathbb{R}$ such that $E(a)=0$ then $E(x)=0 \; \forall x \in \mathbb{R}$. Nor continuity nor other topological properties are needed for this. In exponential fields theory such a function is called a trivial exponential function.

If $E$ is not this trivial function, it's easy to see that we must have $E(0)=1$ and, if $E(x)=1$ for some $x \ne 0$ than $E(x)=1 \forall x \in \mathbb{R}$, i.e. it is a constant (another trivial exp. function).

If $E$ is not trivial and $E(1)=a$ we have one particular funtion $E_a(n)$ and it is easy to see that $E_a(-1)=\dfrac{1}{a}$ and $E_a(n)=a^n$ (and all other properties of integer or fractional exponents), so it is natural to write $E_a(x)=a^x$ .

Now we see that $0^x$ is the null trivial exponential function, and, since this function must be always null, we have $0^0=0$.

  • 2
    @IlmariKaronen: we have: $E(0)=E(a-a)=E(a)E(-a)=0\cdot 0=0\ne 1$. – Emilio Novati Jan 31 '17 at 8:36
  • 2
    True, I missed the fact that the domain of $E$ includes negative numbers. Although it still seems to me that identifying the trivial function $E_0(x)=0\ \forall x\in\mathbb R$ with $0^x$ doesn't really work, because it would also imply that $\frac10=0^{-1}=E_0(-1)=0$. – Ilmari Karonen Jan 31 '17 at 16:35
  • I would argue that $0^x$ is only defined when $x\ge 0$; you would expect $0^-1=\infty$. The property $0^{x+y}=0^x0^y$ holds for all $x,y\ge 0$ whether you define $0^0=0$ or $0^0=1$. – Mike Earnest Sep 14 at 0:33

Take a look at WolframMathWorld's [1] discussion.

See if this gives you any clarification.

[1] Weisstein, Eric W. "Indeterminate." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/Indeterminate.html

  • 3
    @TylerClark: that discussion is about $\lim\limits_{(x,y)\to(a,b)}f(x,y)$ for various $f$ and $(a,b)$. Since that limit does not exist for $f(x,y)=x^y$ at $(x,y)=(0,0)$, $0^0$ is called an indeterminate form. However, that is a separate issue from whether $0^0$ is defined. – robjohn Oct 13 '14 at 0:02
  • 6
    @wendy.krieger: $x^y$ is definitely not continuous at $(0,0)$: $\lim\limits_{x\to0^+}x^0=1$, whereas $\lim\limits_{x\to0^+}0^{\large x}=0$. – robjohn Oct 13 '14 at 5:14
  • 1
    @wendy.krieger: Suppose $x^y$ were continuous at $(0,0)$ and $\lim\limits_{(x,y)\to(0,0)}x^y=1$. Then by definition, $$\forall\epsilon\gt0, \exists\delta\gt0:|(x,y)|\le\delta\implies|x^y-1|\le\epsilon\tag{1}$$ However, $$\forall\delta\gt0,|(0,\delta)|\le\delta\land|0^\delta-1|\gt\tfrac12\tag{2}$$ and $(2)$ contradicts $(1)$. Thus, either $x^y$ is not continuous at $(0,0)$ or $\lim\limits_{(x,y)\to(0,0)}x^y\ne1$. – robjohn Oct 13 '14 at 8:52
  • 3
    @wendy.krieger: Surely, you see that $$\lim_{\delta\to0}\left(\delta^{1/\delta}\right)^{\delta} =\lim_{\delta\to0}\delta=0$$ and that $$\lim_{\delta\to0}\left(\delta^{1/\delta},\delta\right)=(0,0)$$ Thus, along the path $(x,y)=\left(\delta^{1/\delta},\delta\right)$, which tends to $(0,0)$, it is easy to see that $x^y$ tends to $0$. Nowhere is $0$ divided by $0$ or is $0$ raised to any power. Only positive numbers to positive powers. – robjohn Oct 14 '14 at 13:35
  • 3
    @wendy.krieger: if $(x,y)=\left(\delta^{1/\delta},\delta\right)$, then $x=\delta^{1/\delta}$ tends to $0$ and $y=\delta$ tends to $0$. Also $x^y=\left(\delta^{1/\delta}\right)^\delta=\delta$ also tends to $0$. Nothing dodgy. – robjohn Oct 15 '14 at 2:27

A paper for general public is published on Scribd : " Zero to the Zero-th Power" (pp.7-11) : http://www.scribd.com/JJacquelin/documents

  • The graphs and limits in your document are only relevant if one accepts the "continuity rule": if $f$ is discontinuous at a point $p$ then one should not define $f$ at $p$. Without that rule it doesn't make sense to look at graphs and limits when there are other ways to obtain a value. – Mark Feb 12 '17 at 22:05

Source: Understanding Exponents: Why does 0^0 = 1? (BetterExplained article)

A useful analogy to explain the exponent operator of the form $a \cdot b^c$ is to make $a$ grow at the rate $b$ for time $c$.

Expanding on that analogy, $0^0$ can be interpreted as $1\cdot0^0$ which is to say: grow $1$ at the rate of $0$ for time $0$. Since there is no growth (time is $0$), there is no change in the $1$ and the answer is $0^0=1$

Of course, this is just to grok and get an intuition or a feel for it. Science is provisional and so is math in certain areas. 0^0=1 is not always the most useful or relevant value at all times.

Using limits or calculus or binomial theorems doesn't really give you an intuition of why this is so, but I hope this post made you understand why it is so and make you feel it from your spleen.

A clear and intuitive answer can be provided by ZFC Set-Theory. As described in Enderton's 'Elements of Set Theory (available free for viewing here; see pdf-page 151): http://sistemas.fciencias.unam.mx/~lokylog/images/stories/Alexandria/Teoria%20de%20Conjuntos%20Basicos/Enderton%20H.B_Elements%20of%20Set%20Theory.pdf, the set of all functions from the empty set to the empty set consists merely of the empty function which is 1 function. Hence $0^0$ = 1.

Contents taken and reformatted from: The Math Forum

What is $0$ to the $0$ power?

This answer is adapted from an entry in the sci.math Frequently Asked Questions file, which is Copyright (c) 1994 Hans de Vreught (hdev@cp.tn.tudelft.nl). According to some Calculus textbooks, $0^0$ is an "indeterminate form." What mathematicians mean by "indeterminate form" is that in some cases we think about it as having one value, and in other cases we think about it as having another.

When evaluating a limit of the form $0^0$, you need to know that limits of that form are "indeterminate forms," and that you need to use a special technique such as L'Hopital's rule to evaluate them. For instance, when evaluating the limit $\sin(x)^x$ (which is $1$ as $x$ goes to $0$), we say it is equal to $x^x$ (since $\sin(x)$ and $x$ go to $0$ at the same rate, i.e. limit as $x\to0$ of $\sin(x)/x$ is $1$). Then we can see from the graph of $x^x$ that its limit is $1$.

Other than the times when we want it to be indeterminate, $0^0 = 1$ seems to be the most useful choice for $0^0$. This convention allows us to extend definitions in different areas of mathematics that would otherwise require treating $0$ as a special case. Notice that $0^0$ is a discontinuity of the function $f(x,y) = x^y$, because no matter what number you assign to $0^0$, you can't make $x^y$ continuous at $(0,0)$, since the limit along the line $x=0$ is $0$, and the limit along the line $y=0$ is $1$.

This means that depending on the context where $0^0$ occurs, you might wish to substitute it with $1$, indeterminate or undefined/nonexistent.

Some people feel that giving a value to a function with an essential discontinuity at a point, such as $x^y$ at $(0,0)$, is an inelegant patch and should not be done. Others point out correctly that in mathematics, usefulness and consistency are very important, and that under these parameters $0^0 = 1$ is the natural choice.

The following is a list of reasons why $0^0$ should be $1$.

Rotando & Korn show that if $f$ and $g$ are real functions that vanish at the origin and are analytic at $0$ (infinitely differentiable is not sufficient), then $f(x)^{g(x)}$ approaches $1$ as $x$ approaches $0$ from the right.

From Concrete Mathematics p.162 (R. Graham, D. Knuth, O. Patashnik):

Some textbooks leave the quantity 0^0 undefined, because the functions $0^x$ and $x^0$ have different limiting values when $x$ decreases to $0$. But this is a mistake. We must define $x^0=1$ for all $x$, if the binomial theorem is to be valid when $x=0$, $y=0$, and/or $x=-y$. The theorem is too important to be arbitrarily restricted! By contrast, the function $0^x$ is quite unimportant.

Published by Addison-Wesley, 2nd printing Dec, 1988.

As a rule of thumb, one can say that $0^0 = 1$, but $0.0^{0.0}$ is undefined, meaning that when approaching from a different direction there is no clearly predetermined value to assign to $0.0^{0.0}$ ; but Kahan has argued that $0.0^{0.0}$ should be $1$, because if $f(x)$, $g(x)$ $\to0$ as $x$ approaches some limit, and $f(x)$ and $g(x)$ are analytic functions, then $f(x)^{g(x)}\to1$.

The discussion of $0^0$ is very old. Euler argues for $0^0 = 1$ since $a^0 = 1$ for a not equal to $0$. The controversy raged throughout the nineteenth century, but was mainly conducted in the pages of the lesser journals: Grunert's Archiv and Schlomilch's Zeitshrift. Consensus has recently been built around setting the value of $0^0 = 1$.

References

Knuth. Two notes on notation. (AMM 99 no. 5 (May 1992), 403-422).

H. E. Vaughan. The expression '$0^0$'. Mathematics Teacher 63 (1970), pp.111-112.

Louis M. Rotando and Henry Korn. The Indeterminate Form $0^0$. Mathematics Magazine, Vol. 50, No. 1 (January 1977), pp. 41-42.

L. J. Paige. A note on indeterminate forms. American Mathematical Monthly, 61 (1954), 189-190; reprinted in the Mathematical Association of America's 1969 volume, Selected Papers on Calculus, pp. 210-211.

Baxley & Hayashi. A note on indeterminate forms. American Mathematical Monthly, 85 (1978), pp. 484-486.

Robert S. Fouch. On the definability of zero to the power zero. School Science and Mathematics 53, No. 9 (December 1953), pp. 693-696.

Let $$p=\lim\limits_{x\to 0} x^x$$

Then $$ln(p)=\lim\limits_{x\to 0} x\ln(x)$$

$$ln(p)=\lim\limits_{x\to 0} \frac{\ln(x)}{x^{-1}}$$

We can use L'Hopital on RHS as both numerator and denominator is tending to $\infty$

$$ln(p)=\lim\limits_{x\to 0} -\frac{1/x}{x^{-2}}$$

This can be simplified by cancelling some x terms out.

$$ln(p)=\lim\limits_{x\to 0} -x$$

Thereafter the limit can be applied

$$ln(p)=0$$

$$ p = e^0$$

Therefore $$p=1$$

That's a nice question, and not an easy one. Actually, I don't think anybody has written a definite solution (if there is, it's probably some naive based on intuition); but if you accept Russell's way (which is identical to that of Cantor in respect to natural numbers), take a look at Whitehead's paper http://www.jstor.org/stable/2370026. In the third line of page 23 (page 388 on the paper) he defines the concept of number (cardinal) exponentiation. Recurring to Cantor's set theory is the only way I think this problem could be solved. The use of real analysis, or other branches such as combinatorics (binomial theorem), won't do any good because their construction already pressuposes cardinal arithmetic. I.e, the problem can be solved using a more elementary theory.

Have a nice day.

The longstanding practice of leaving $0^0$ undefined is usually justified with arguments based on path dependent limits on the real numbers. As we see here, however, it is also possible to justify this practice based on purely discrete methods.

If the intuition of exponentiation on $N$ (where $0\in N$) is to be repeated multiplication such that $x^2=x\cdot x$, then we can formally justify the following definition:

  1. $\forall x,y\in N: x^y\in N$ (a binary function on $N$)

  2. $\forall x\in N:(x\ne 0\implies x^0=1)$

  3. $\forall x,y\in N:x^{y+1}=x^y\cdot x$

Here, $0^0$ is assumed to be a natural number, but no specific value is assigned to it.

From this definition, we can derive the usual Laws of Exopnents:

  1. $\forall x,y,z\in N: (x\ne 0 \implies x^y \cdot x^z = x^{x+y})$

  2. $\forall x,y,z\in N: (x\ne 0 \implies (x^y)^z = x^{y\cdot z})$

  3. $\forall x,y,z\in N: (x,y\ne 0 \implies (x\cdot y)^z = x^z\cdot y^z)$

For a detailed development based on formal proofs, see "Oh, the Ambiguity!" at my math blog.

Follow-up

A better approach, I have since found, is to construct a partial function on $N\times N$ with domain of definition $N\times N \setminus (0,0)$ such that:

  1. $\forall x,y:[x,y\in N \land \neg [x=0\land y=0] \implies x^y\in N$

  2. $\forall x:[x\in N\land x\ne 0\implies x^0=1]$

  3. $0^1=0$

  4. $\forall x,y:[x,y\in N \land \neg [x=0\land y=0] \implies x^{y+1}=x^y\cdot x]$

See my revised blog posting at above link (originally dated October 9, 2013).

  • 6
    It seems forced to decide that $0^0$ is "undefined", there is no contradiction in having $0^0=1$, and in fact it makes the second axiom, as well all the laws to change from implication to just $x^0=1$ and $x^{y+1}=x^y\cdot x$, and so on. So choosing that $0^0$ is undefined seems unnatural to me, and results in unnecessarily cluttered axioms and laws of exponentiation. – Asaf Karagila Nov 20 '13 at 21:37
  • 5
    This seems to compare to a situation where I'll write "Let's agree that the cardinality of the empty set is not defined, now we can devise the usual axioms of cardinal arithmetics, but I'll have to add an implication of the form only if the sets involved are not empty ... to every axiom!" – Asaf Karagila Nov 20 '13 at 21:39
  • @AsafKaragila There is also no contradiction for $0^0=999$ or any other natural number. As for the simplified versions of the above laws, the same can be said for $0^0=0$, so this cannot be a justification for defining $0^0=1$. $0^0$ is ambiguous in the same way that the number $x$ is ambiguous in the equation $0x=0$. Any value will work, as I show at my blog. I am not aware of any logically compelling reason to choose any particular value. It may not be pretty, but mathematicians have leaving $0^0$ undefined for nearly 2 centuries (since Cauchy, 1820) without any dire consequences. – Dan Christensen Nov 20 '13 at 21:59
  • 1
    There's also no contradiction in deciding $x^y=999$ for every $x,y$. So what? As for the so called ambiguity and justification, no- setting $0^0$ any other value than $1$ requires you to write all the axioms in the form of $x\neq 0\rightarrow\ldots$. Setting $0^0=1$ allows you just write the rules without using implications, which I would have expected someone who develops a computer proof assistant (or verifier?) to appreciate as a way of reducing complexity of statements. Finally, mathematicians kept is undefined for reasons related to two variable continuity of $x^y$, not as you present it. – Asaf Karagila Nov 20 '13 at 22:08
  • 1
    $0^0=999$ would be a contradiction to the power laws, because then $(0^0)^2 = 999^2 \ne 0^{0\cdot2} = 999$. The only two values for $0^0$ consistent with the power laws are $0$ and $1$. – celtschk Jul 11 '15 at 17:03

There are several ways to prove that $0^0$ := 1. One way is to define the function of 2 variables f(x,y) = $x^y$. Then, evaluate the limit $\lim_{(x,y)\to(0,0)}x^y$. You can see - as I did it with Matlab - that $x^y$ goes to 1 when x $\rightarrow$ 0 and y $\rightarrow$ 0. Be aware that we cannot say $x^y$ equals 1 when x and y goes to 0, but $x^y$ approaches 1 as x and y goes to 0. enter image description here

  • 1
    Your claim $x^y \to 1$ as $(x,y) \to (0,0)$ involves uncountably infinitely many $(x,y)$, so you can't prove it by listing finitely many MATLAB output. – GNUSupporter 8964民主女神 地下教會 May 17 at 11:12
  • 2
    Given that the Question here is over seven years old, a quick post that contains numerical experiments but no mathematical reasoning will not advance the understanding provided by the two dozen previous Answers. – hardmath May 17 at 11:20

We must note that the term $0^0$ is not a normal number since there is no arithmetic calculation that leads to form the $0^0$. We have to options:

A) We set $0^0=0$ and demand that all arithmetic operations are subjected to number $0$ and not the term $0^0$, e.g. writing $0^0=0^{(1-1)}$ is forbidden. This definition is nice since adding zero doesn't change any equation.

$\space$

B) We set $0^0=1$ and demand that all arithmetic operations are subjected to number $1$ and not the term $0^0$, e.g. writing $0^0=0^{(1-1)}$ is forbidden. This definition is nice since the limit $x^0 \rightarrow 1$ as $x \rightarrow 0$ is consistent with this definition.

Summa summarum: if we want to define the number $0^0$, then the most useful choice is $1$.

It is solvable by means of contradiction:

Here are some general rules:

$x^0=1$

and

$0^x=0$

If $x=0$ you get two answers for $0^0$: $0,1$

It can't be two different solutions so we simply coin it as undefined.

  • 5
    The second "general rule" is wrong. It is correct only for positive numbers. – Anixx Aug 21 '15 at 2:58
  • 1
    The argument contains a second error "not ($P$ and $Q$) $\Longrightarrow$ (not $P$) and (not $Q$)". In other words, if two statements $0^0=0$ and $0^0=1$ contradict each other, then both are wrong! – Mark Feb 12 '17 at 21:51

protected by user99914 May 22 at 0:13

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

Not the answer you're looking for? Browse other questions tagged or ask your own question.