Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

If we have given that U is a functor $ D \to C $ creates limits and C is Complete. and we are to show that D is complete and U preserves limits.

I follow its in this way

As we know that U is a functor $ D \to C $ creates limits if and only if U reflects limits that means all the diagrams D $ I \to D $ and all cones ($ q_X $ : $ A \to DX )$ where $X \in I$ on D. if $ Uq_X : UA \to UDX $ is a limit cone on D. This implies that UD has a limit in C, hence C has all limits. so it is complete.

However, I don't understand how we can prove D is complete by using this definition, if it satisfies above.

share|cite|improve this question
Actually creating and reflecting limits are different things as shown in either this link or this other link. – Giorgio Mossa Jan 6 '14 at 19:43
up vote 0 down vote accepted

The statement $U \colon \mathbf D \to \mathbf C$ create limits means that for every functor/diagram $F \colon \mathbf J \to \mathbf D$ for every limit cone $\langle \tau_j \colon d \to U \circ F(j)\rangle_{j \in \mathbf J}$ there's a unique limit cone $\langle \sigma_j \colon c \to F(j)\rangle_{j \in \mathbf J}$ such that $U(\sigma_j)=\tau_j$ for every $j \in \mathbf J$.

If $\mathbf C$ was complete (I suppose you mean small complete) then for every category (small) $\mathbf J$ and functor $F' \colon \mathbf J \to \mathbf C$ we have a limit cone $\langle \tau_j \colon d \to F'(j)\rangle_{j \in \mathbf J}$.

This in particular implies that for every functor $F \colon \mathbf J \to \mathbf D$ the functor $U \circ F \colon \mathbf J \to \mathbf C$ has a limit cone $\langle \tau_j \colon d \to U\circ F(j)\rangle_{j \in \mathbf J}$. By the property of $U$ of creating limit there must exists also a limit cone $\langle \sigma_j \colon c \to F(j)\rangle_{j \in \mathbf J}$: hence $F$ has a limit in $\mathbf D$.

To prove that $U$ preserve limits let suppose that $\langle \sigma_j \colon c \to F(j)\rangle_{j \in \mathbf J}$ is a limit cone in $\mathbf C$ for $F$.

If $\langle \tau_j \colon d \to U\circ F(j)\rangle_{j \in \mathbf J}$ was limit cone for $U \circ F$ then by what we have seen above there's $\langle \sigma'_j \colon c' \to F(j)\rangle_{j \in \mathbf J}$ which is a limit for $F$. But since also $\sigma$ was a limit cone for $F$ there must be an isomorphism $i \colon c \to c'$ in $\mathbf C$ that is a isomorphism of cones $f \colon \sigma \to \sigma'$ (i.e. for every $j \in \mathbf J$ we have that $\sigma'_j \circ i = \sigma_j$). $U(i)$ must is still an isomorphism between the cones $U(\sigma_j)$ and $U(\sigma')=\tau$, and since $\tau$ was a limit cone then also $U(\sigma)$ is a limit cone for $U \circ F$. So $U$ preserve limits.

Hope this helps.

share|cite|improve this answer

If $D$ is a diagram in $\cal D$, then $UD$ is a diagram in $\cal C$. By completeness, there is a limiting cone $(a,\ \alpha_k:a\to UDk)_{k\in J}$. Now $U$ creates limits, so there is a unique cone $(b,\ \beta_k:b\to Dk)_{k\in J}$, such that $Ub=a$ and $U\beta_k=\alpha_k$, and this cone is a limiting cone in $\cal D$. This shows that $\cal D$ is complete.
If $(b',\ \beta'_k:b'\to Dk)_{k\in J}$ is another limiting cone over $D$, then there is a isomorphism $f:b\to b'$ such that $β'_kf=β_k$. But then $Uf$ is an isomorphism $a\to Ub'$ such that $Uβ'_k\circ Uf=α_k$, which means that $(Ub', Uβ'_k:Ub'\to UDk)_{k\in J}$ is also a limiting cone over $UD$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.