Update recursive-datatypes.md

some spelling corrections

microsite/src/main/tut/interpreters/recursive-datatypes.md

@@ -14,7 +14,7 @@ It is quite easy to define schemas for recursive data types, but extra care must
 
 ## Building interpreters
 
-The construction of an interpreter is guarantied to terminate under certain conditions. The `SelfReference` case of the `SchemaF` ADT has a lazy `unroll` field that "performs the recursion". Evaluating this field eagerly in an algebra will make the interpretation diverge, but keeping it unevaluated will guaranty the interpretation to terminate.
+The construction of an interpreter is guarantied to terminate under certain conditions. The `SelfReference` case of the `SchemaF` ADT has a lazy `unroll` field that "performs the recursion". Evaluating this field eagerly in an algebra will make the interpretation diverge, but keeping it unevaluated will guarantee the interpretation to terminate.
 
 For example, if we were to write an interpreter for `org.scalacheck.Gen` (that is, to implement an `HAlgebra[Schema, Gen]`), writing the following makes the process diverge:
 
@@ -38,7 +38,7 @@ But if we keep the `unroll` field unevaluated, by passing it as a by-name parame
 
 ## Non-termination of the derived functor
 
-It is impossible to guaranty non-termination of a derived functor in general, but here are a few tips.
+It is impossible to guarantee non-termination of a derived functor in general, but here are a few tips.
 
 * Contravariant functors resembling to a function `A => X` (like, for example, `play.api.libs.json.Writes`) are safe as long as the input data is acyclic.
 * Covariant functors resembling to a function `X => A` (like `play.api.libs.json.Reads`) are generaly safe, as long as you can prove that "each step consumes a bit of the `X`".