The handshaking lemma for $2$-graphs says that for any simple graph $G$ $$ 2e(G) = \sum_{v \in G} d(v). $$
Is there a generalization of this fact to $r$-uniform hypergraphs for $r>2$ ?
Let's compute the degree sum and see what happens: $$ \sum_{v \in G} d(v) = \sum_{v \in G} |\{e \in E(G) \ : \ v \in e \}| $$ How many times are we counting each edge? Well, each edge is being counted $r$ times, once for each $v \in e$ (not sure about this reasoning here). Thus we must have that $$ \frac{1}{r}\sum_{v \in G} d(v) = e(G) $$
Here is a possible justification for this using an auxillary bipartite graph. Let $H$ be a bipartite graph with partitions $A,B$ such that $A = V(G)$ and $B = E(G)$ with adjacency defined as follows: for each $v \in A$ and $e \in B$ we have $$ v \sim e \iff v \in e. $$ Counting the edges of $H$ from the perspective of each partition gives $$ r\cdot e(G)= \sum_{b \in B}d_H(b) = e(H) = \sum_{a\in A} d_H(a) = \sum_{v \in V(G)} d_G(v). $$
(Slightly meta follow up question: If my reasoning is correct then why is this more generalized form of the handshaking lemma not publicized more? Maybe my search queries are not articulated well enough.)
Copy and pasted from original post:
Let's compute the degree sum and see what happens: $$ \sum_{v \in G} d(v) = \sum_{v \in G} |\{e \in E(G) \ : \ v \in e \}| $$ How many times are we counting each edge? Well, each edge is being counted $r$ times, once for each $v \in e$ (not sure about this reasoning here). Thus we must have that $$ \frac{1}{r}\sum_{v \in G} d(v) = e(G) $$
Here is a possible justification for this using an auxillary bipartite graph. Let $H$ be a bipartite graph with partitions $A,B$ such that $A = V(G)$ and $B = E(G)$ with adjacency defined as follows: for each $v \in A$ and $e \in B$ we have $$ v \sim e \iff v \in e. $$ Counting the edges of $H$ from the perspective of each partition gives $$ r\cdot e(G)= \sum_{b \in B}d_H(b) = e(H) = \sum_{a\in A} d_H(a) = \sum_{v \in V(G)} d_G(v). $$