Given a $0$-connected $H$-space $X$, we can prove that left and right multiplication are (weak) homotopy equivalences. This allows us to construct the Hopf fibrations for $S^1$, $S^3$ and $S^7$. I always thought that $S^0$ was the odd one out due to it not being connected however after reading page 4 of "H-spaces from a homotopy point of view" I see that it is claimed $0$-connectedness can be weakened to $\pi_0(X)$ being a group.

Does this mean the real Hopf fibration can be constructed using the Hopf construction?

If so could I have some modern references?

This question originally stemmed from the discussion here on the nforum.


This is completely elementary. From the H-space structure $X\times X\to X$ you construct the Hopf fibration $X*X\to SX$ via $(x,t,y)\to (xy,t)$, where you think of $SX$ as the quotient of $X\times I$ by identifying both the $t=1$- and $t=-1$-level to one point, respectively. For $X=S^0$ you get a fibration $S^1\to S^1$, which you „see“ is the map of degree $2$: each point has exactly two preimages. So it is actually the fibration $S^1\to RP^1$.

| cite | improve this answer | |
  • $\begingroup$ What are the necessary conditions for $X*X \to SX$ to be a fibration? I seem to recall X being path connected is used some essential way. $\endgroup$ – Ali Caglayan Feb 11 '19 at 18:26
  • 1
    $\begingroup$ For $X=S^0$ you just see ad hoc what the map is: it is a 2-fold covering. $\endgroup$ – ThiKu Feb 11 '19 at 19:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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