Is there a generic way to memoize in Scala?

Class/trait level val compiles to a combination of a method and a private variable. Hence a recursive definition is allowed.

Local vals on the other hand are just regular variables, and thus recursive definition is not allowed.

By the way, even if the def you defined worked, it wouldn't do what you expect. On every invocation of foo a new function object fib will be created and it will have its own backing map. What you should be doing instead is this (if you really want a def to be your public interface):

private val fib: Memo[Int, BigInt] = Memo {
  case 0 => 0
  case 1 => 1
  case n => fib(n-1) + fib(n-2) 
}

def foo(n: Int) = {
  fib(n)
} 

Scalaz has a solution for that, why not reuse it?

import scalaz.Memo
lazy val fib: Int => BigInt = Memo.mutableHashMapMemo {
  case 0 => 0
  case 1 => 1
  case n => fib(n-2) + fib(n-1)
}

You can read more about memoization in Scalaz.

I found a better way to memoize using Scala:

def memoize[I, O](f: I => O): I => O = new mutable.HashMap[I, O]() {
  override def apply(key: I) = getOrElseUpdate(key, f(key))
}

Now you can write fibonacci as follows:

lazy val fib: Int => BigInt = memoize {
  case 0 => 0
  case 1 => 1
  case n => fib(n-1) + fib(n-2)
}

Here's one with multiple arguments (the choose function):

lazy val c: ((Int, Int)) => BigInt = memoize {
  case (_, 0) => 1
  case (n, r) if r > n/2 => c(n, n - r)
  case (n, r) => c(n - 1, r - 1) + c(n - 1, r)
}

And here's the subset sum problem:

// is there a subset of s which has sum = t
def isSubsetSumAchievable(s: Vector[Int], t: Int) = {
  // f is (i, j) => Boolean i.e. can the first i elements of s add up to j
  lazy val f: ((Int, Int)) => Boolean = memoize {
    case (_, 0) => true        // 0 can always be achieved using empty list
    case (0, _) => false       // we can never achieve non-zero if we have empty list
    case (i, j) => 
      val k = i - 1            // try the kth element
      f(k, j - s(k)) || f(k, j)
  }
  f(s.length, t)
}

EDIT: As discussed below, here is a thread-safe version

def memoize[I, O](f: I => O): I => O = new mutable.HashMap[I, O]() {self =>
  override def apply(key: I) = self.synchronized(getOrElseUpdate(key, f(key)))
}