Multiple alignment

Multiple alignment points with no gap between expressions is obtained using the alignat environment from amsmath. The following MWE illustrates that:

\documentclass{article}

\usepackage{amsmath}

\begin{document}

\begin{alignat*}{3}
  &[x \mapsto s]x &&= s && \\
  &[x \mapsto s]y &&= y \qquad &&\text{als } y\neq x \\
  &[x \mapsto s](\lambda(y)t_1) &&= \lambda(y)[x \mapsto s]t_1 \qquad &&\text{als } y \neq x \text{ en   } y \not \in FV(s) \\
  &[x \mapsto s](t_1 \; t_2)&&= ([x \mapsto s]t_1)\;([x \mapsto s]t_2) &&
\end{alignat*}

\end{document}

I've taken the liberty to convert the unknown \la{<arg>} macro to \lambda(<arg>).

Consider reading Herbert's mathmode document. It illustrates the various "alignment environments" of amsmath clearly. Specifically, see section 25 align environments, p 43 onward.

using just align would be done like this:

\documentclass{article}
\usepackage{amsmath}% http://ctan.org/pkg/amsmath
\begin{document}
\begin{align*}
  [x \mapsto s]x &= s \\
  [x \mapsto s]y &= y &&\text{als } y\neq x \\
  [x \mapsto s](\lambda(y)t_1) &= \lambda(y)[x \mapsto s]t_1 &&\text{als } y \neq x \text{ en   } y \not \in FV(s) \\
  [x \mapsto s](t_1 \; t_2) &= (x \mapsto s]t_1)\;([x \mapsto s]t_2)
\end{align*}
\end{document}

this gives the following result: