Ordered Pairs

The ordered pair is not an isolated curiosity — it is the load-bearing joint connecting sets to nearly every other mathematical structure. A relation between sets $A$ and $B$ is officially defined as a set of ordered pairs, i.e., a subset of $A \times B$ (the Cartesian product); a function is a relation satisfying an extra uniqueness condition on second coordinates. This means that concepts as basic as "$f(x)=y$" are, underneath, statements about membership of a Kuratowski pair $\{\{x\},\{x,y\}\}$ in a set of pairs — though no one thinks about it this way in daily practice, the fact that it is possible to think about it this way is what guarantees the whole apparatus of functions and relations rests on a single consistent foundation. In computer science, ordered pairs underlie the implementation of data structures: a `cons` cell in Lisp-family languages, a key–value entry in a dictionary or hash map, and a labeled edge in a directed graph are all conceptually ordered pairs, and formal semantics for programming languages frequently model them exactly as Kuratowski-style nested sets or their type-theoretic analogues. In cryptography and coding theory, ordered pairs appear as coordinates in finite fields and lattices used to build ciphers. In statistics, a bivariate data point $(x_i, y_i)$ used in scatterplots and regression is precisely an ordered pair, and the entire theory of joint distributions is built on ordered tuples of random variables. In physics and engineering, coordinate pairs $(x,y)$ in the plane, phase-space pairs $(q,p)$ of position and momentum, and complex numbers represented as pairs $(a,b)$ for $a+bi$ all rely on the same underlying guarantee: that swapping coordinates genuinely changes the object. Finally, in pure mathematics, the recursive definition of $n$-tuples ($(a,b,c) := ((a,b),c)$, and so on) and the construction of Cartesian products, sequences, and matrices all bottom out in the single theorem proved here.

Ask the tutor: Can you show me step by step why $(a,b) = (c,d)$ forces $a=c$ before it forces $b=d$, using a concrete example where $a \ne b$?

Think of an unordered pair $\{a,b\}$ as a photograph of two people standing side by side with no label saying who is "first." An ordered pair needs a way to break that symmetry using only set-builder notation — no arrows, no coordinates, no new primitive concept of "sequence." Kuratowski's idea is to smuggle in an asymmetry through nesting depth and membership, not through any new kind of object. The set $\{a\}$ inside $(a,b) = \{\{a\},\{a,b\}\}$ acts like a flag: it is the one-element set among the two elements of $(a,b)$ (assuming $a \ne b$), and its single member tells you unambiguously which of $a,b$ came first. The other element, $\{a,b\}$, is a two-element set containing both coordinates, so once you know the singleton reveals the first coordinate, subtracting it from $\{a,b\}$ (via a symmetric difference, or just "the other one") reveals the second. When $a = b$, the pair degenerates to $\{\{a\}\}$, a single-element set, which still unambiguously recovers $a$ as both coordinates. The proof of the central theorem makes this recovery procedure completely rigorous. ```svg ```

```svg ```

Example 1: Verifying $(2,7) \ne (7,2)$ from the raw definition, and locating the discrepancy explicitly. By definition,

Note first that $\{2,7\} = \{7,2\}$ as unordered sets (extensionality: both contain exactly the elements $2$ and $7$). So we can rewrite:

These two sets share the element $\{2,7\}$. Do they share the other element? We ask: is $\{2\} = \{7\}$? By extensionality this would require $2 = 7$, which is false. So $\{2\} \ne \{7\}$, hence the sets $\{\{2\},\{2,7\}\}$ and $\{\{7\},\{2,7\}\}$ differ in one of their two elements and are therefore not equal (again by extensionality: $\{2\}$ is an element of $(2,7)$ but is not an element of $(7,2)$, since the elements of $(7,2)$ are $\{7\}$ and $\{2,7\}$, neither of which equals $\{2\}$ — checking: $\{7\}\ne\{2\}$ since $7\ne2$, and $\{2,7\}\ne\{2\}$ since $\{2,7\}$ has two elements while $\{2\}$ has one). Hence

This matches the theorem: $(2,7)=(7,2)$ would require $2=7$ and $7=2$ simultaneously (taking $a=2,b=7,c=7,d=2$ in the biconditional), which fails, so the theorem correctly predicts inequality without our needing to unpack sets at all — but it is instructive to see the raw computation agree.

Example 2: The degenerate case $a=b$, worked with asymmetric numbers to show the mechanism explicitly — computing $(9,9)$ and checking it is not accidentally equal to $(9,4)$. By definition,

using $\{9,9\}=\{9\}$ (a set is unaffected by listing an element twice — this is exactly the extensionality principle that sets are determined solely by which elements they contain, not by repetition or order in a listing). So $(9,9)$ is a genuinely one-element set, its only element being the singleton $\{9\}$. Now compare with

This is a two-element set (since $\{9\}\ne\{9,4\}$, because $4 \in \{9,4\}$ but $4\notin\{9\}$). A one-element set can never equal a two-element set: if $\{\{9\}\} = \{\{9\},\{9,4\}\}$, then in particular $\{9,4\}$, being an element of the right-hand side, would have to be an element of the left-hand side too (by extensionality, equal sets have identical elements); but the only element of $\{\{9\}\}$ is $\{9\}$, so we would need $\{9,4\} = \{9\}$, which is false since $4 \ne 9$ means $4$ is a member of the left set but not the right. Contradiction. Hence

exactly as the theorem predicts, since it would require $9=4$ in the second coordinate. This example also illustrates the necessity of the case split in the proof: had we tried to apply the "generic" Case 2 argument (which assumes $a\ne b$) directly to $(9,9)$, the step "$\{a,b\}$ has two distinct elements" would have silently failed, which is exactly why the proof isolates $a=b$ as its own case.