A student argues: 'Since rationals are dense in the reals, if I pick a real number uniformly at random, it's more likely to be rational than irrational.' What is wrong with this reasoning?
ANothing — density means rationals are everywhere, so the probability should be high
BDensity is a topological property about neighborhoods, not a measure-theoretic one; the rationals have measure zero and a randomly chosen real is rational with probability zero
CThe argument is correct for small intervals but fails for the whole real line
DRationals are actually not dense in the reals, so the premise is false
Density (every open interval contains a rational) and measure (how much of the line rationals 'take up') are different properties. The rationals are countable, so in Lebesgue measure they occupy zero length on the real line. Despite being topologically dense, they are measure-theoretically negligible — a randomly chosen real is irrational with probability 1. This tension between density and measure is one of analysis's most important lessons.
Question 2 Multiple Choice
In the proof that a rational exists between any two reals a < b, which tool is used to find a denominator q large enough that consecutive multiples 1/q, 2/q, 3/q, ... are spaced more finely than the interval (a,b)?
AThe completeness of ℝ (every Cauchy sequence converges)
BThe uncountability of the irrationals
CThe Archimedean property (for any ε > 0, there exists n ∈ ℕ with 1/n < ε)
DThe intermediate value theorem
The key step is finding q with 1/q < b − a, ensuring the grid of multiples p/q is finer than the gap between a and b. That some such q exists is exactly what the Archimedean property guarantees: for any positive real (here, b − a), there exists a natural number n with 1/n below it. Once this denominator exists, the smallest integer p with p/q > a satisfies p/q < a + 1/q < a + (b−a) = b.
Question 3 True / False
Between any two distinct real numbers, there are infinitely many rational numbers.
TTrue
FFalse
Answer: True
Once you know one rational r exists in (a, b), you can apply the same argument to the sub-interval (a, r) to find another, and repeat without end. More directly, if p/q lies in (a,b), so do p/(q+1), p/(q+2), ... for sufficiently large denominators (though careful checking is needed). The core result — every open interval contains infinitely many rationals — is a direct consequence of density.
Question 4 True / False
Because the rationals are dense in ℝ, they make up a positive fraction of the length of any interval on the real line.
TTrue
FFalse
Answer: False
The rationals are countable, and countable sets have Lebesgue measure zero. A set of measure zero is, in the measure-theoretic sense, negligible: you could cover it with open intervals of total length as small as you like. Density tells you that no point is 'far' from a rational (topological statement), but measure tells you that rationals collectively have no width (metric statement). These properties coexist without contradiction — the rationals are simultaneously everywhere-close and measure-zero.
Question 5 Short Answer
Explain the apparent paradox: the rationals are dense in ℝ (every real number is arbitrarily close to rationals) and yet, in a precise measure-theoretic sense, 'almost all' real numbers are irrational.
Think about your answer, then reveal below.
Model answer: Density is a topological statement about proximity: every open neighborhood of every real number contains rationals. Measure is a statement about size: the total 'length' of the rationals on the real line is zero. These are different properties of different mathematical structures (topology vs. measure theory). The rationals are a countable set — they can be listed in a sequence — and any countable set can be covered by open intervals of arbitrarily small total length, so its measure is zero. Density says you can always find a rational nearby; measure zero says that if you chose a real number truly at random, the probability of hitting a rational is exactly zero.
Analogy: imagine placing individual points (zero area each) throughout the plane. No matter how many you place, if they are countable, they cover zero area — but they can still be dense. The rationals are like a very fine but zero-area lattice threaded through ℝ. The irrationals fill in all the genuine 'length.'