How to interpret correlation functions in QFT?
Yes, in scalar field theory, $\langle 0 | T\{\phi(y) \phi(x)\} | 0 \rangle$ is the amplitude for a particle to propagate from $x$ to $y$. There are caveats to this, because not all QFTs admit particle interpretations, but for massive scalar fields with at most moderately strong interactions, it's correct. Applying the operator $\phi({\bf x},t)$ to the vacuum $|0\rangle$ puts the QFT into the state $|\delta_{\bf x},t \rangle$, where there's a single particle whose wave function at time $t$ is the delta-function supported at ${\bf x}$. If $x$ comes later than $y$, the number $\langle 0 | \phi({\bf x},t)\phi({\bf y},t') | 0 \rangle$ is just the inner product of $| \delta_{\bf x},t \rangle$ with $| \delta_{\bf y},t' \rangle$.
However, the function $f(x,y) = \langle 0 | T\{\phi(y) \phi(x)\} | 0 \rangle$ is not actually a correlation function in the standard statistical sense. It can't be; it's not even real-valued. However, it is a close cousin of an honest-to-goodness correlation function.
If make the substitution $t=-i\tau$, you'll turn the action $$iS = i\int dtd{\bf x} \{\phi(x)\Box\phi(x) - V(\phi(x))\}$$ of scalar field theory on $\mathbb{R}^{d,1}$ into an energy function $$-E(\phi) = -\int d\tau d{\bf x} \{\phi(x)\Delta\phi(x) + V(\phi(x))\}$$ which is defined on scalar fields living on $\mathbb{R}^{d+1}$. Likewise, the oscillating Feynman integral $\int \mathcal{D}\phi e^{iS(\phi)}$ becomes a Gibbs measure $\int \mathcal{D}\phi e^{-E(\phi)}$.
The Gibbs measure is a probability measure on the set of classical scalar fields on $\mathbb{R}^{d+1}$. It has correlation functions $g(({\bf x}, \tau),({\bf y},\tau')) = E[\phi({\bf x}, \tau)\phi({\bf y},\tau')]$. These correlation functions have the property that they may be analytically continued to complex values of $\tau$ having the form $\tau = e^{i\theta}t$ with $\theta \in [0,\pi/2]$. If we take $\tau$ as far as we can, setting it equal to $i t$, we obtain the Minkowski-signature "correlation functions" $f(x,y) = g(({\bf x},it),({\bf y},it'))$.
So $f$ isn't really a correlation function, but it's the boundary value of the analytic continuation of a correlation function. But that takes a long time to say, so the terminology gets abused.
No, $⟨0|T{ϕ(y)ϕ(x)}|0⟩$ is NOT the probability amplitude for a particle to propagate from $x$ to $y$, even for a free scalar field. It seems to be a common false belief that it is. There is one obvious reason and one deep reason why it cannot be.
The obvious reason is that the square of this value, which is supposed to be the probability density, does not integrate to 1 (see wikipedia):
$⟨0|T{ϕ(y)ϕ(x)}|0⟩=$ $$G_F(x,y) =\lim_{\epsilon\to 0}\frac{1}{(2\pi)^4}\int d^4p\frac{e^{-ip(x-y)}}{p^2-m^2+i\epsilon} =\begin{cases} -\frac{1}{4\pi}\delta(s)+\frac{m}{8\pi\sqrt{s}}H_1^{(1)}(m\sqrt{s}) & s\ge 0\\ -\frac{im}{4\pi^2\sqrt{-s}}K_1(m\sqrt{-s}) & s<0 \end{cases}$$ where $s:=(x^0-y^0)^2-(\vec{x}-\vec{y})^2$, hence $\int dy_1 dy_2 dy_3\,|⟨0|T{ϕ(y)ϕ(x)}|0⟩|^2$ is infinite (if makes sense at all). Interpretation as a 'relative probability amplitude' does not fix that because the most of 'probability to propagate from $x$ to $y$' would be anyway concentrated on the cone $s=0$ due to $\delta(s)$-term.
A more deep reason is that 'probability' does not make any sense until a probability space, i.e. the set of possible outcomes of an experiment, is introduced. The least we have to assume is that these outcomes are mutually exclusive. In nonrelativistic quantum mechanics this set is an orthonormal basis in a Hilbert space (or a bit more general object such as the set of delta-functions supported at different spatial points $\mathbf{y}$). Then the inner product with the basis elements is interpreted as the transition probability density. But for the free quantum field, delta-functions supported at different spatial points $\mathbf{y}$ and the same time $t$ are not orthogonal anymore (in no sense): $⟨0|T{ϕ(\mathbf{y},t)ϕ(\mathbf{x},t)}|0⟩\ne 0$ by the above expression. In other words, one can find the particle at different points simultaneously, making just impossible to speak of the probability density to find a particle at a point. Hence, the inner product of $|δ_x,t⟩$ with $|δ_y,t'⟩$ cannot be interpreted as probability density (keeping aside the question what does this notation at all mean in the Fock space), quite opposite to the answer by user1504.
Finally, one cannot introduce any experiment to measure the quantity under name 'the probability for a free particle to propagate from $x$ to $y$', because the precision of measurement of the particle coordinates cannot be better than the particle Compton wavelength.
Beware: having absolutely no understanding of the free quantum field, writing just to get a chance to be corrected by those who have. In particular, would be grateful for explanation what is meaning of $⟨0|T{ϕ(y)ϕ(x)}|0⟩$, if it is not a probability amplitude. Being stuck with this question for quite a long time.