It has been said that the Big Bang started from a singularity. Think about a balloon radially growing over time. Fix a time $t_0, t_1 > 0$, and let $M_0, M_1$ be two balloons at time $t_0, t_1$ respectively. I can find a two-parameter diffeomorphism $\phi(t_0, t_1): M_0 \rightarrow M_1$.

However, I cannot find a diffeomorphism if I let $t_0 = 0$ and $t_1 > 0$, i.e. $\phi(0, t_1): \{*\} \rightarrow M_1$. In what sense should I interpret a homotopy between initial state (big bang) and final state (the current universe)? Is it even true that the Big Bang started from a singularity?

  • 13
    $\begingroup$ Are you equally disturbed by the existence of a cone, where the circles at two different heights are diffeomorphic except at height zero. $\endgroup$
    – WillO
    May 15 at 19:42
  • 2
    $\begingroup$ @willO In a sense, yes. I am very inclined to work with smooth manifold (even though "sharp corners" can be replaced out by some smooth approximations.) $\endgroup$
    – James C
    May 16 at 19:07
  • 1
    $\begingroup$ Since 2010, there's been a cosmological model which "avoids the singularity" required by General Relativity: It's "Cosmology with torsion", & can be found at arxiv.org/a/poplawski_n_1.html . It uses Einstein-Cartan Theory (based on a 1929 collaboration between Einstein and the mathematician Ely Cartan), instead of 1915's GR. (The subtitle "An alternative to cosmic inflation", on the 2010 paper, is a bit misleading: Poplawski's model's inflationary, but does not require hypothetical "inflaton" particles. It's also past- & future-eternal, which is contrary to some belief systems.) $\endgroup$
    – Edouard
    May 17 at 3:16
  • $\begingroup$ The big bang theory isn't about the "moment of beginning" of the universe, but about its very early phase: m.youtube.com/watch?v=gpgor5hdD8s $\endgroup$
    – M. Winter
    May 17 at 8:41

5 Answers 5


The singularity at the start of the universe in the Big Bang model is not supposed to be understood as part of the smooth manifold of spacetime, precisely for this reason.

The time function on spacetime does not actually assign a "point" to $t = 0$. It's undefined (otherwise spacetime wouldn't be a Lorentzian manifold), and the same is true if you take an FLRW universe and try to keep the initial spatial slice 3d - since the scale factor goes to zero, the manifold is not Lorentzian there.

If you want to model the initial singularity of the Big Bang as part of spacetime, you need to consider more general models of spacetime than a Lorentzian manifold.


A “singularity” is shorthand for “a place where our current understanding of the laws of physics breaks down”. If we use our current model of the laws of physics (specifically general relativity) to extrapolate the state of the early universe backwards in time, we find that density and temperature are unbounded - loosely speaking, we reach a state of infinite density and temperature. This is not physically possible. Therefore something else must happen - but at the moment we are not sure exactly what that is. It seems likely that we will need to have a theory of quantum gravity to explain the very first moments of the universe.

  • 8
    $\begingroup$ As physicist, I have had always problem with the sentence "where the laws of physics (...) break down". For me it is a model of the considered physical process that fails and not the laws of physics. If someone talks about infinite density and temperature, like in en.wikipedia.org/wiki/Big_Bang, ref. 20, I wish I could ask him how he precisely defines density and temperature in that case. $\endgroup$ May 15 at 15:34
  • 7
    $\begingroup$ @JanGogolin I agree that the laws of physics themselves are not at fault. Note that I actually said "where the laws of physics as we currently understand them break down". An alternative way of putting this would be "where our current model of the laws of physics breaks down". I also agree that talking about infinite physical quantities is imprecise. I have amended my wording in both places. $\endgroup$
    – gandalf61
    May 15 at 15:52
  • $\begingroup$ I have clearly missed the meaning of your amendments. My comments were directed more towards the popular explanation as for example in wikipedia and in ref. 20 there. $\endgroup$ May 15 at 17:44
  • $\begingroup$ @JanGogolin The "laws of physics" are a model. Reality has broken them frequently when we've discovered a deficiency in the laws. $\endgroup$
    – OrangeDog
    May 16 at 14:39
  • 3
    $\begingroup$ @JanGogolin: This is semantics. Either A) the "laws of physics" are an imperfect model of reality, which is (supposed to be) accurate to within experimental error, over the limited domain in which we have actually tested it, or B) the "laws of physics" are some kind of Platonic ideal, and perfectly correct. But in case (B), the "laws of physics" can only be determined indirectly through experiment and observation, which are subject to measurement error etc., so regardless of which definition you use, the "laws" that we print in textbooks must be of type (A). $\endgroup$
    – Kevin
    May 17 at 2:15

You're asking a mathematical question. Mathematics is a human invention. Sometimes it maps well onto the phenomena we see. Spacetime geometry is a case where, in the portion of the universe accessible to observation, the math and the phenomena are in decent agreement.

Part of this is interpreting the observations as indicating that the universe was once in a much hotter and denser state than it is at the present time. That's really what we mean by the "big bang". But we really don't know enough physics to reliably extrapolate back to the "beginning", whatever that was.

Mathematical models of physical phenomena are prone to singularities, but when the theoretical singularities should be accessible to observation, they are never seen. This is a limitation of using math to model the universe.


There are several problems with this argument.

First, cosmological models having a singularity just means that they blow up as $t\to 0^+$ and aren't defined at $t=0$. That's interpreted as a limitation of the models, not evidence that real-world quantities were infinite at $t=0$. Several other answers have already covered this.

Second, even inasmuch as you can complete the topological manifold and treat the model as having a $t=0$, the completion usually isn't a single point, because no single point can be in the causal past of the whole space at early $t$ (this is the horizon problem). Even if $k>0$, meaning the space at every $t>0$ has a finite volume and it goes to zero at $t\to 0$, the completion is still topologically a 3-sphere, not a point.

Third, even if you could complete the manifold with a single point, that isn't obviously a problem. If you put polar coordinates on Euclidean space, there is no homotopy from the space at $r=0$ to the space at any $r\ne 0$. To be fair, I don't think anything closely analogous to that Euclidean example can happen with Lorentzian signature, for the handwaving reason that at the central point all directions would have to be timelike, which is inconsistent with the assumed signature. But to prove that you have to make a coordinate-independent argument, not an argument that depends on a $t$ coordinate and associated foliation.

I think there are (not widely accepted) models in which the universe starts with a Euclidean signature, so that $t=0$ can be a mere coordinate singularity like the $r$ of the previous paragraph. I don't know how the transition from one signature to another is dealt with.


Physicists feel a great aversion to infinity-valued physical quantities. They call them singularities and usually say that the physics fails there. A mathematician would never say the mathematics fails, because the value of $1/x$ function at $x=0$ is infinite. The inverse of this function is clearly defined and equal to zero there.

Coming to your example: imagine a two dimensional universe with constant positive Gaussian curvature $K$. It is two-dimensional manifold called the two-sphere ($S^2$). There are only two intrinsic properties characterizing it, the Gaussian curvature $K$ and the total surface area $A$. They are not independent, but related to each by Gauss-Bonnet theorem:

\begin{equation} \int_{A}~K~dA~=~A\cdot K~=~4\pi \end{equation}

Due to this equation, in case of the surface area A going to zero, the curvature K diverges to infinity. However, their product remains finite being $4\pi$. Moreover, $S^n$-spheres are not contractible to a point (they have no interior). A sphere with vanishing surface area remains topologically a sphere despite of its infinite curvature.

Your picture with balloon parametrized by the time I would replace with a fiber bundle $R^{+}\times S^2$ with fiber $S^2$ and base its surface area $A$. In conclusion, the places with infinite values of some physical quantities are interesting ones and not just abhorrence, or a proof that the underlying theory has its limit there. See for example Is the Planck force a truly "Planck unit"?.

  • 5
    $\begingroup$ Actually, we have few aversions to singularities in models. We throw around things like Dirac δ in ways that profoundly annoy fussy mathematicians. However, we never see actual singularities in experiments. Thus, we presume that singularities in the math belong to the math, not the physics. $\endgroup$
    – John Doty
    May 14 at 17:19
  • 2
    $\begingroup$ @JohnDoty, you are of course right, we do not see singularities in experiments, to me just because there are none in our physical world. I do not like using the word "singularity" .For me it means only that we have not quite understood what happens there. See for example physics.stackexchange.com/a/707944/281096 . $\endgroup$ May 14 at 17:42

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