Unit 3, lesson 40

Product and Quotient Properties of Logarithms

Why the logarithm of a product is a sum and the logarithm of a quotient is a difference, when these rules hold, and how to expand and condense logarithms without changing the domain by accident.

  • Theory part 79
  • Practice part 80
  • Properties of Logarithms

Two years of doubling followed by three more years of doubling make five years of doubling, because 22⋅23=252^2 \cdot 2^3 = 2^5. When powers of the same base are multiplied, their exponents add. Logarithms read off exponents, so they turn products into sums and quotients into differences. Before calculators, this is how long multiplications were done, by adding numbers looked up in tables of logarithms.

Key terms

Logarithm of a product

Definition Let b>0b > 0 with b≠1b \ne 1, and let M>0M > 0 and N>0N > 0. Then

log⁡b(MN)=log⁡bM+log⁡bN.\log_b(MN) = \log_b M + \log_b N.

To prove it, put u=log⁡bMu = \log_b M and v=log⁡bNv = \log_b N, so that bu=Mb^u = M and bv=Nb^v = N. Then MN=bubv=bu+vMN = b^u b^v = b^{u + v}, and by the definition of the logarithm log⁡b(MN)=u+v\log_b(MN) = u + v.

In plain words The logarithm of a product is the sum of the logarithms. It is the rule “multiply powers, add exponents” read through logarithms.

Example Since 8⋅32=2568 \cdot 32 = 256,

log⁡2(8⋅32)=log⁡2256=8andlog⁡28+log⁡232=3+5=8.\begin{gathered} \log_2(8 \cdot 32) = \log_2 256 = 8 \\[0.4em] \text{and} \\[0.4em] \log_2 8 + \log_2 32 = 3 + 5 = 8. \end{gathered}

The rule also works when the separate logarithms are not nice numbers. The values log⁡2≈0.30103\log 2 \approx 0.30103 and log⁡5≈0.69897\log 5 \approx 0.69897 add up to exactly log⁡10=1\log 10 = 1.

Logarithm of a quotient

Definition Let b>0b > 0 with b≠1b \ne 1, and let M>0M > 0 and N>0N > 0. Then

log⁡b(MN)=log⁡bM−log⁡bN.\log_b\left(\frac{M}{N}\right) = \log_b M - \log_b N.

With u=log⁡bMu = \log_b M and v=log⁡bNv = \log_b N as before,

MN=bubv=bu−v,\frac{M}{N} = \frac{b^u}{b^v} = b^{u - v},

so the logarithm of the quotient is u−vu - v. Taking M=1M = 1 gives the special case log⁡b1N=−log⁡bN\log_b \frac{1}{N} = -\log_b N.

In plain words The logarithm of a quotient is the logarithm of the top minus the logarithm of the bottom. It is not the logarithm of the top divided by the logarithm of the bottom.

Example Since 2439=27\frac{243}{9} = 27,

log⁡3(2439)=log⁡327=3andlog⁡3243−log⁡39=5−2=3.\begin{gathered} \log_3\left(\frac{243}{9}\right) = \log_3 27 = 3 \\[0.4em] \text{and} \\[0.4em] \log_3 243 - \log_3 9 = 5 - 2 = 3. \end{gathered}

Dividing the logarithms instead gives 52\frac{5}{2}, which is wrong.

Positive arguments

Definition The product and quotient rules require each of MM and NN to be positive, not only their product or quotient. If M<0M < 0 and N<0N < 0, then MN>0MN > 0, so log⁡b(MN)\log_b(MN) is defined while log⁡bM\log_b M and log⁡bN\log_b N are not. For any real MM and NN with MN>0MN > 0, which is the same as MN>0\frac{M}{N} > 0, the correct statements are

log⁡b(MN)=log⁡b∣M∣+log⁡b∣N∣,log⁡b(MN)=log⁡b∣M∣−log⁡b∣N∣.\begin{gathered} \log_b(MN) = \log_b \lvert M \rvert + \log_b \lvert N \rvert, \\[0.8em] \log_b\left(\frac{M}{N}\right) = \log_b \lvert M \rvert - \log_b \lvert N \rvert. \end{gathered}

In plain words Before you split a logarithm, check that every piece you write has a positive argument. Splitting can shrink the set of inputs where the expression makes sense.

Example The expression log⁡2((−4)(−8))\log_2\big((-4)(-8)\big) is fine and equals log⁡232=5\log_2 32 = 5, but log⁡2(−4)+log⁡2(−8)\log_2(-4) + \log_2(-8) is undefined. The version with absolute values gives log⁡24+log⁡28=2+3=5\log_2 4 + \log_2 8 = 2 + 3 = 5. With a variable, the expression log⁡((x+1)(x−3))\log\big((x + 1)(x - 3)\big) is defined when x<−1x < -1 or x>3x > 3, while log⁡(x+1)+log⁡(x−3)\log(x + 1) + \log(x - 3) is defined only when x>3x > 3. At x=−2x = -2 the first equals log⁡5\log 5 and the second is undefined.

Common base

Definition The product and quotient rules combine logarithms that have the same base bb. A sum log⁡bM+log⁡cN\log_b M + \log_c N with b≠cb \ne c is in general not equal to log⁡b(MN)\log_b(MN) or to log⁡c(MN)\log_c(MN). To combine logarithms with different bases, first rewrite them with one base, which the change of base formula in the next lesson does.

In plain words You can only merge logarithms that have the same base. The common logarithm log⁡\log and the natural logarithm ln⁡\ln have different bases, so they do not merge either.

Example The sum

log⁡28+log⁡48=3+32=92,\log_2 8 + \log_4 8 = 3 + \frac{3}{2} = \frac{9}{2},

since 43/2=84^{3/2} = 8. Merging the arguments gives 6464, but log⁡264=6\log_2 64 = 6 and log⁡464=3\log_4 64 = 3, and neither is 92\frac{9}{2}.

Expanding

Definition Expanding a logarithm means using the product and quotient rules to rewrite the logarithm of a product or quotient as a sum and difference of logarithms of its factors. Each step is valid only where every new argument is positive, so an expanded form comes with a condition on the variables.

In plain words Expanding breaks one complicated logarithm into several simple ones. In calculus this makes derivatives of products and quotients much easier, a method called logarithmic differentiation.

Example For x>0x > 0 and y>0y > 0,

log⁡5(25xy)=log⁡525+log⁡5x−log⁡5y=2+log⁡5x−log⁡5y.\begin{aligned} &\log_5\left(\frac{25x}{y}\right) \\[0.8em] &\quad = \log_5 25 + \log_5 x - \log_5 y \\[0.4em] &\quad = 2 + \log_5 x - \log_5 y. \end{aligned}

For a single variable, consider ln⁡x(x+1)x−2\ln \frac{x(x + 1)}{x - 2}. When x>2x > 2 all three factors are positive, so

ln⁡x(x+1)x−2=ln⁡x+ln⁡(x+1)−ln⁡(x−2),x>2.\begin{gathered} \begin{aligned} &\ln \frac{x(x + 1)}{x - 2} \\[0.8em] &\quad = \ln x + \ln(x + 1) - \ln(x - 2), \end{aligned} \\[0.4em] x > 2. \end{gathered}

The left side is also defined for −1<x<0-1 < x < 0, where the quotient is positive, but there ln⁡x\ln x and ln⁡(x−2)\ln(x - 2) are undefined. The equation holds only for x>2x > 2.

Condensing

Definition Condensing means using the rules from right to left to combine sums and differences of logarithms with the same base into a single logarithm. For M,N,P>0M, N, P > 0,

log⁡bM+log⁡bN−log⁡bP=log⁡bMNP.\begin{aligned} &\log_b M + \log_b N - \log_b P \\[0.4em] &\quad = \log_b \frac{MN}{P}. \end{aligned}

The condensed expression may be defined for more values of the variable than the original one, so keep track of the original domain.

In plain words Condensing packs several logarithms into one, which is the usual first step in solving an equation with logarithms.

Example With numbers,

log⁡64+log⁡69=log⁡636=2.\log_6 4 + \log_6 9 = \log_6 36 = 2.

With a variable, for x>0x > 0

ln⁡x+ln⁡(x+3)−ln⁡2=ln⁡x(x+3)2.\begin{aligned} &\ln x + \ln(x + 3) - \ln 2 \\[0.4em] &\quad = \ln \frac{x(x + 3)}{2}. \end{aligned}

The left side needs x>0x > 0, while the right side is also defined for x<−3x < -3. If an equation leads to a value like x=−4x = -4, the condensed form accepts it but the original expression does not, so that value is not a solution.

Sums do not split into logarithms

Definition There is no rule that splits the logarithm of a sum or a difference. For positive MM and NN,

log⁡b(M+N)=log⁡bM+log⁡bN\log_b(M + N) = \log_b M + \log_b N

holds exactly when M+N=MNM + N = MN, because the right side equals log⁡b(MN)\log_b(MN) and a logarithm is one-to-one. In the same way log⁡b(M−N)\log_b(M - N) is in general not log⁡bM−log⁡bN\log_b M - \log_b N. What does work is factoring the sum first, for example

log⁡b(M+N)=log⁡b(M(1+NM))=log⁡bM+log⁡b(1+NM).\begin{aligned} &\log_b(M + N) \\[0.4em] &\quad = \log_b\left(M\left(1 + \frac{N}{M}\right)\right) \\[0.8em] &\quad = \log_b M + \log_b\left(1 + \frac{N}{M}\right). \end{aligned}

In plain words Logarithms turn products into sums. They do not turn sums into anything simpler. If you see a sum inside a logarithm, try to factor it.

Example A quick numerical check exposes the mistake. The value of log⁡(10+10)\log(10 + 10) is log⁡20≈1.30103\log 20 \approx 1.30103, but log⁡10+log⁡10=2\log 10 + \log 10 = 2. When the sum factors, the product rule applies to the factors. For x>0x > 0

ln⁡(x2+5x)=ln⁡(x(x+5))=ln⁡x+ln⁡(x+5).\begin{aligned} \ln\left(x^2 + 5x\right) &= \ln\big(x(x + 5)\big) \\[0.4em] &= \ln x + \ln(x + 5). \end{aligned}

Check yourself

  1. Write log⁡42+log⁡432\log_4 2 + \log_4 32 as a single logarithm and evaluate it.
Show the answer

By the product rule the sum is log⁡464\log_4 64, and 43=644^3 = 64, so the value is 33.

  1. Expand log⁡39xx+4\log_3 \frac{9x}{x + 4} for x>0x > 0.
Show the answer

For x>0x > 0 the factors 99, xx and x+4x + 4 are positive, so

log⁡39xx+4=2+log⁡3x−log⁡3(x+4).\begin{aligned} &\log_3 \frac{9x}{x + 4} \\[0.8em] &\quad = 2 + \log_3 x - \log_3(x + 4). \end{aligned}

The left side is also defined for x<−4x < -4, but the expanded form is not.

  1. Solve log⁡2x+log⁡2(x−2)=3\log_2 x + \log_2(x - 2) = 3.
Show the answer

Condensing gives log⁡2(x(x−2))=3\log_2\big(x(x - 2)\big) = 3, so x(x−2)=8x(x - 2) = 8, which means x2−2x−8=0x^2 - 2x - 8 = 0 and x=4x = 4 or x=−2x = -2. The original equation needs x>2x > 2, so x=−2x = -2 is rejected. The only solution is x=4x = 4, and indeed log⁡24+log⁡22=2+1=3\log_2 4 + \log_2 2 = 2 + 1 = 3.

How this lesson connects

Where this leads in Calculus