$$\DeclareMathOperator{\obj}{obj}\ \DeclareMathOperator{\arr}{arr}\ \DeclareMathOperator{\dom}{dom}\ \DeclareMathOperator{\cod}{cod}$$

# 2.1 More Categories

## 2.1.1 Dual Category

Take any category and turn the arrows around, you have another category called the dual category. For a category $$A$$ the dual or opposite category is denoted $$A^{op}$$.

## 2.1.2 Generators and relations

This presentation is a nice trick. Give all the objects and some arrows, and some equations between the arrows. For example given object x and arrow $$e: x \rightarrow x$$ and relation $$e^4 = 1_x$$ the category has arrows $$1_x, e, e^2, e^3$$.

## 2.1.3 Special arrows

Arrows of a category can have special properties.

## 2.1.4 Monic

An arrow $$m:A\rightarrow B$$ is monic if for any pair of arrows $$f_1, f_2: D \rightarrow A$$ then $f_1 \cdot m = f_2 \cdot m \implies f_1 = f_2$ in other words, $$m$$ can be cancelled in post position. In $$\bf Sets$$ injections (one to one functions) are monic.

## 2.1.5 Epi

An arrow $$e:A\rightarrow B$$ is epi if for any pair of arrows $$g_1, b_2: B \rightarrow C$$ then $e \cdot g_1 = e \cdot g_2 \implies g_1 = g_2$ in other words, $$m$$ can be cancelled in pre position. in $$\bf Sets$$ surjections (onto functions) are epi

## 2.1.6 Invertible

An arrow $$f:A\rightarrow B$$ is invertible if there is an arrow $$f^{-1}: B\rightarrow A$$ called the inverse such that $f\cdot f^{-1} = 1_A$ and $f^{-1} \cdot f = 1_B$ In this case the inverse is unique and obviously also invertible, both arrows are called isomorphisms, and $$A$$ and $$B$$ are said to be isomorphic.

## 2.1.7 Universal Algebra

As you can see a lot of information about the structure of a system can be modelled purely with functions. Any model using functions alone is said to be abstract. Category theory is the algebra of functions and thus the theory of abstraction.

Many laws of category theory contains qualifications and read $$P$$ is unique, up to isomorphism. This means it isn't unique at all, but all the things satisfying the condition are isomorphic.

This lack of precision is one of the most difficult things to handle when building programming models: it means we have many choices of representation and theory leaves it open which ones to choose.