My Math Forum Negative exponent to a power

 Elementary Math Fractions, Percentages, Word Problems, Equations, Inequations, Factorization, Expansion

 July 26th, 2019, 07:03 AM #1 Newbie   Joined: Jul 2019 From: Florida Posts: 2 Thanks: 0 Negative exponent to a power A practice test I'm using to study has a question that raises a fraction to a negative exponent which is also raise to a (positive) exponent. To me, the simple answer would be: (1/4)^-4^2 = (4^-1)^16 but the book says the 16 is negative. Any reason why this is? The book only gives what I've written above (with a -16 expoent) as the explanation so I'm a little bit in the dark. I'll keep working on this and remove the post if I figure it out.
 July 26th, 2019, 08:06 AM #2 Senior Member     Joined: Sep 2015 From: USA Posts: 2,529 Thanks: 1389 your expression is $\left(\left(\dfrac 1 4\right)^{-4}\right)^2$ ? I would do this as $\left(\left(\dfrac 1 4\right)^{-4}\right)^2 = \\ \left(\dfrac 1 4\right)^{-8} = \\ 4^8 = \\ 65536$
 July 26th, 2019, 08:08 AM #3 Senior Member   Joined: Jun 2019 From: USA Posts: 120 Thanks: 40 Did they write the problem with carets the way you did? It may be an ambiguity issue, as it is not clear whether they mean $\displaystyle (1/4)^{((-4)^2)}=4^{-16}$ or $\displaystyle (1/4)^{(-(4^2))}=4^{16}$ or $\displaystyle ((1/4)^{-4})^2=4^{8}$. I would normally agree with the first answer, but I know programming languages that would interpret it as one of the second two.
July 26th, 2019, 08:19 AM   #4
Senior Member

Joined: Aug 2012

Posts: 2,355
Thanks: 737

Exponentiation has higher precedence (or binds more tightly, in computer lingo) than addition/subtraction. So the expression $-4^2$ means: first do $4^2$, which is $16$, then apply the $-$ to get $-16$.

The fact that some (Which? I'm curious) programming languages get that wrong is irrelevant. The math rules are clear and unambiguous. I blame the math teachers of the world for failing to properly teach order of precedence. That's why these awful "This question broke the Internet" things go around that are nothing more than simple order of precedence problems.

FWIW I just put -4**2 into the interactive Python interpreter and got -16, the correct answer.

ps -- Perl also, correct evaluation.

Quote:
 Originally Posted by Questionairre I'll keep working on this and remove the post if I figure it out.

Last edited by Maschke; July 26th, 2019 at 08:27 AM.

July 26th, 2019, 08:29 AM   #5
Senior Member

Joined: Jun 2019
From: USA

Posts: 120
Thanks: 40

Quote:
 Originally Posted by Maschke The fact that some (which?) programming languages get that wrong is irrelevant. The math rules are clear and unambiguous.
I agree, but then how did the practise test get 4^16?

I just tried MATLAB and Excel, and they both gave me 65536.

July 26th, 2019, 08:41 AM   #6
Senior Member

Joined: Aug 2012

Posts: 2,355
Thanks: 737

Quote:
 Originally Posted by DarnItJimImAnEngineer I agree, but then how did the practise test get 4^16? I just tried MATLAB and Excel, and they both gave me 65536.
$(\frac{1}{4})^{-16} = \frac{1}{(\frac{1}{4})^{16}}= 4^{16} = 4294967296$. Your post has me confused. Excel and MATLAB are clearly completely wrong, if you entered this correctly.

Oh wait I see what they did. Exponentiation binds right to left, $a^{b^c}$ means $a^{(b^c)}$. That's why exponential towers (tetration) grow so quickly. MATLAB and Excel are clearly doing it the wrong way. They're computing $16^2$. That is wrong and there is no question about it.

I just tried it in C, but C has no exponentiation operator. It uses a pow() function in math.h, so you can't screw it up. If you write pow(-4,2) that's 16, and - pow(4,2) is -16. I hope I don't spend the rest of the day tracking down which languages get this wrong.

Last edited by Maschke; July 26th, 2019 at 09:02 AM.

July 26th, 2019, 09:00 AM   #7
Senior Member

Joined: Jun 2019
From: USA

Posts: 120
Thanks: 40

Quote:
 Originally Posted by Maschke Excel and MATLAB are clearly completely wrong, if you entered this correctly.
They both use left-to-right priority for order of operations with exponents, just like they do with multiplication/division and addition/subtraction.

I think I may have had a lysdexic attack or two somewhere in this post. Ultimately, though, I think the question is this:
Does order of operations consider $\displaystyle -4^2$ to be $\displaystyle 0-4^2=-16$ or "(negative four) squared equals 16?"

July 26th, 2019, 09:07 AM   #8
Senior Member

Joined: Aug 2012

Posts: 2,355
Thanks: 737

tl;dr: Wiki says that exponential towers always bind right to left, but iterated ^'s SOMETIMES bind left to right, and they mention Excel and MATLAB as two examples. I'll leave this as I originally wrote it.

Quote:
 Originally Posted by DarnItJimImAnEngineer They both use left-to-right priority for order of operations with exponents, just like they do with multiplication/division and addition/subtraction.
That's completely wrong for exponentiation. It's not ambiguous or controversial. They're wrong. I imagine Excel screwed this up in 1980-something and for backward compatibility they realized they can never fix it. I don't know what MATLAB's excuse is, that program is used in a lot of scientific work, isn't it? $2^{3^4} = 2^{81}$ and there is simply no question about it.

Quote:
 Originally Posted by DarnItJimImAnEngineer I think I may have had a lysdexic attack or two somewhere in this post. Ultimately, though, I think the question is this: Does order of operations consider $\displaystyle -4^2$ to be $\displaystyle 0-4^2=-16$ or "(negative four) squared equals 16?"
Exponentiation binds more tightly than subtraction so $-4^2 = 0 - 4^2 = -16$ and there is no question about that either. If you want $(-4)^2 = 16$ you have to use the parens.

I hate to sound so dogmatic, but I'm stating the mathematical conventions.

Excel I don't mind, because it's used mostly for business. But MATLAB is a scientific tool. They have no excuse for getting this wrong.

I'm sounding pretty dogmatic even to myself. Am I missing something? Wolfram Alpha certainly gets it right.

https://www.wolframalpha.com/input/?i=(1%2F4)%5E-4%5E2

There are those who say this question is controversial or ambiguous. I do not believe that to be the case. The mathematical order of precedence is absolutely standard as I've stated it. I am genuinely shocked that MATLAB could possibly interpret an iterated exponential left to right. That's wrong. That would make tetration wrong. I'm not too surprised that Excel is wrong, it was designed as a business tool and now they're stuck with the wrong convention.

ps -- Mea culpa, mea culpa, mea maxima culpa. Here is what Wikipedia says. I think they're wrong but they're Wikipedia and I'm not.

If exponentiation is indicated by stacked symbols using superscript notation, the usual rule is to work from the top down ...

However, when using operator notation with a caret (^) or arrow (â†‘), there is no common standard.[13] For example, Microsoft Excel and computation programming language MATLAB evaluate a^b^c as $(a^b)^c$, but Google Search and Wolfram Alpha as $a^{(b^c)}$. Thus 4^3^2 is evaluated to 4,096 in the first case and to 262,144 in the second case.

I'm genuinely surprised. I have always regarded iterated ^'s as binding right to left. Frankly I still believe MATLAB and Excel are wrong and so is Wiki, but clearly the weight of authority is on the other side. Although Wolfram is THE authority. I truly believe MATLAB and Excel are wrong, but at least Wiki has noted the confusion.

pps -- Finally, here is that [13] reference. Someone went to the trouble to survey a whole bunch of programming languages on both the iterated ^ and the -4^2 issues, and what do you know, there really is a lot of inconsistency out there.

https://codeplea.com/exponentiation-...tivity-options

In math, the conventions are as I stated them. But there's much more variance among programming languages, database languages, and spreadsheets than I realized.

Last edited by Maschke; July 26th, 2019 at 10:01 AM.

 July 26th, 2019, 09:32 AM #9 Newbie   Joined: Jul 2019 From: Florida Posts: 2 Thanks: 0 OP Wow, okay this blew up way more than I thought. Sorry for not using everyones username in this response to the entirety of the thread so far. This was actually an algebra problem I'm doing on paper, not in a programming language. To the poster mentioning the order of precedence, that took care of all of my issues. To the one who mentioned the math teachers who have failed the youth, I agree. From what I remember in classes, most of the students would also lie and say they understand what the teacher had gone over because they just don't care to learn. It's a strange bilateral failure with neither side understanding what is really going on. I'll leave the problem up in the forum for others to see and learn from as one poster mentioned. Thank you all for you input and feel free to have what ever discourse you'd like related to the problem. This is your thread now.
 July 26th, 2019, 10:47 AM #10 Senior Member   Joined: Jun 2019 From: USA Posts: 120 Thanks: 40 To be fair to MATLAB, I've been doing engineering calculations for a couple decades, and I don't think I've ever had to use nested exponents in real life -- except maybe with e as one of the bases, and then it uses the exp() function. And anytime I have expressions that complicated, I always add more parentheses than are strictly necessary just for human readability. The negation point was one I had a real doubt about, though. I suppose I never think of negation as subtraction. But you're right; if WolframAlpha says negate after exponentiation, I'm inclined to trust them as mathematically superior to me. "Also, you should really just use parentheses. If you don't, you're going to have a bad day sooner or later." Dogma and principles aside, I think we can all agree with that one, no?

 Tags exponent, negative, power

 Thread Tools Display Modes Linear Mode

 Similar Threads Thread Thread Starter Forum Replies Last Post Angus Math Software 9 July 23rd, 2017 04:07 AM Abhimanyu Elementary Math 4 October 4th, 2015 10:06 PM Toyboy Algebra 4 July 30th, 2013 04:16 AM prashantakerkar Algebra 13 November 16th, 2011 08:14 AM cytotoxictcell Algebra 1 November 1st, 2011 11:43 AM

 Contact - Home - Forums - Cryptocurrency Forum - Top