1

This question refers to an example illustrating properties of the Yoneda lemma: Basic Example of Yoneda Lemma?

Here the paragraph of matter in my question:

enter image description here

By Definition the Yoneda lemma provides a bijection between $Hom(Hom_{C}(-,A), F) $ and $F(A)$, so in this case $Hom(Hom_{M}(-,*), \phi) \cong \phi(*)$.

How do we conclude here that YL provedes the identification of a $M$-action $\phi$ (defined above as functor $M \to Set$) with a function $h: M \to \phi(*)$ with the rule $h(m_1m_2) = \phi(m_1)(h(m_2))$? I don't see where does it come.

user557
  • 12,359
user267839
  • 9,217
  • What is $A^*$? Not sure if it matters, but it is a bit confusing. – Thomas Andrews Oct 03 '17 at 23:53
  • It doesn't seem to be true. If and $M$-action is on $A$ and on $B$, then there is an $M$-action on $A\sqcup B$. Which map $M\to A\sqcup B$ is this $h$? – Thomas Andrews Oct 04 '17 at 00:00
  • So if $\eta \in Hom(Hom_{M}(-,), \phi)$ in natural trafo, so you mean that $h = \eta_(id_)$ (by def of Yoneda isomorphism)? Indeed, both a maps $M \to \phi()$. But why $h(m_1m_2) = \phi(m_1)(h(m_2))$ holds? – user267839 Oct 04 '17 at 00:08

1 Answers1

3

It doesn't seem to be true. If there is an $M$-action on $A$ and on $B$, then there is an $M$-action on $A\sqcup B$. Which map $M\to A\sqcup B$ is this $h$?

The Yoneda lemma here says that each $x\in \phi(\star)$ there is a unique map $h_x:M\to \phi(\star)$ so that $h_x(m_1m_2)=\phi(m_1)h_x(m_2)$. That is, $h_x$ is a natural transformation from $\phi$ to $\hom_{M}(-,*)$. This map is $h_x(m)=\phi(m)(x).$

This in turn means that $\phi$ is equivalent to an $h:M\times X\to X$ with the usual definition of an action: $h(m_1,h(m_2,x))=h(m_1m_2,x).$ and $h(1,x)=x$.

Thomas Andrews
  • 186,215
  • 2
    You are correct. I did make a mistake there and write down not what I intended. The $x$ is needed to even make the correspondence to the right fold work. Thanks. – Derek Elkins left SE Oct 04 '17 at 00:18