Laws of logarithms
In a minute
The turn a product inside a log into a sum, a quotient into a difference and a power into a multiple. In AI you use them with base and base to simplify expressions, to turn exponential and power models into straight lines, and to compare values on log scales such as decibels.
You'll be able to
- Use the product, quotient and power laws to split one into several, or to combine several into one.
- Say when the laws apply: positive values inside every log, base or in exams, and no law for the log of a sum.
- Take or of an exponential or power model to get a straight line, and undo the logs to recover the model.
- Use the laws to compare values on a log scale and to solve log equations, rejecting any value that makes a log undefined.
Before you start2 quick questions on what this builds on
Revise first: 1.5 Exponents and introductory logarithms
The three laws
A is an exponent. because , and because .
Multiply the two numbers: . By the , multiplying powers adds the exponents. So .
So the log of a product is the sum of the logs. Dividing powers subtracts exponents, and a power of a power multiplies them. These give the other two laws.
These are the . They hold when , and are positive and . They are in the formula booklet; you need to spot when to use them.
In AI exams the base is always or . Base is the , so the product law reads .
Since and , the power law gives and .
Try it first, then reveal one step at a time.
- Use the power law first, so every term is a single log: .
- Adding logs multiplies the numbers inside: .
- Subtracting logs divides them: .
- GDC check: and , so they agree.
Try it first, then reveal one step at a time.
- Quotient law: .
- Product law on the first term: .
- , since . Power law: .
- So the expression is . The factor inside became outside: multiplying by adds to a base- log.
Straight lines from exponential and power models
Many AI models are exponential, , or power, . Their graphs curve, but taking logs of both sides turns each into a straight line.
Exponential model: . That is the product law, then the power law with .
So against is a straight line, with gradient and vertical intercept .
Power model: . So against is a straight line, with gradient and intercept . It works with too.
To undo it, run the laws backwards: combine the right-hand side into one log, then drop the logs from both sides. You'll use this to fit models to data in 2.10.
Try it first, then reveal one step at a time.
- Take of both sides: .
- Product law: .
- Power law, with : . So , a straight line in .
- Interpret: the gradient is added to each hour, so is multiplied by each hour, about growth per hour.
Try it first, then reveal one step at a time.
- Power law, read backwards: .
- Write the constant as a log: .
- Product law, read backwards: . Equal logs mean equal numbers, so , with and .
- Substitute , keeping exact:
- So the fish has a mass of about kg. Since , doubling the length multiplies the mass by .
Log scales and log equations
Sound level in decibels (dB) is . Here is the sound's intensity and is a fixed reference intensity.
Multiply the intensity by . The product law gives . So the level rises by dB, whatever it was to start with.
Try it first, then reveal one step at a time.
- Five hairdryers give times the intensity, so .
- The level rises by dB.
- New level: , so dB to 1 d.p.
- Interpret: five times the intensity adds only about dB, because decibels measure the log of the intensity.
The laws also help you solve equations. Combine the logs on each side into a single log, then drop the logs. Always check each answer in the original equation, because a log of zero or a negative number doesn't exist.
Try it first, then reveal one step at a time.
- Product law on the left: .
- Equal logs mean equal numbers: , so .
- Factorise: , so or .
- Check in the original equation. gives , which doesn't exist, so reject it.
- For : , so is the only solution. Your GDC's solver confirms it.
Common mistakes
- Splitting the log of a sum. can't be split: the laws only work on products, quotients and powers inside the log.
- Applying a power to only part of the inside. , not .
- Treating as . The quotient law is about , the log of a fraction, not a fraction of two logs.
- Rounding the gradient or intercept of a log line, often read from your GDC, before you undo the logs. A small change in an exponent changes the answer a lot, so store full values and round only the final answer.
- In a "show that", substituting values into the given result to see that it works. Working backwards earns no marks: start from the given model and apply the laws until you reach the result.
Exam tips
- "Show that" with logs: start from the given expression, use one law per line and name it. Your final line must match the given result exactly, or the last mark is lost.
- When asked to write an expression in the form , write each step. A correct use of one law can earn a method mark even if a later step slips.
- Check a simplification on your GDC: put the same test value, one where every log is defined, into the original expression and into your answer. The two values must agree; if they don't, the simplification is wrong.
- "Solve" means find every valid solution. With a GDC, write down the equation you give the GDC, then the solution, and say why you rejected any value.
Key takeaways
Practise this
4 past-paper questions on Laws of logarithms in the AI HL bank, each with a mark scheme and a worked explanation.