diff --git a/AUTHORS.md b/AUTHORS.md
index 6fe886fd20..93283a4a7d 100644
--- a/AUTHORS.md
+++ b/AUTHORS.md
@@ -209,6 +209,7 @@
 - Prasasto Adi
 - Safwan Kamarrudin
 - Arthur Tilley
+- Vilhjalmur Thorsteinsson
 
 ## Others whose work we've taken and included in NLTK, but who didn't directly contribute it:
 ### Contributors to the Porter Stemmer
diff --git a/nltk/probability.py b/nltk/probability.py
old mode 100644
new mode 100755
index 423eb7272e..bc5a2bd98f
--- a/nltk/probability.py
+++ b/nltk/probability.py
@@ -105,6 +105,9 @@ def __init__(self, samples=None):
         """
         Counter.__init__(self, samples)
 
+        # Cached number of samples in this FreqDist
+        self._N = None
+
     def N(self):
         """
         Return the total number of sample outcomes that have been
@@ -114,7 +117,38 @@ def N(self):
 
         :rtype: int
         """
-        return sum(self.values())
+        if self._N is None:
+            # Not already cached, or cache has been invalidated
+            self._N = sum(self.values())
+        return self._N
+
+    def __setitem__(self, key, val):
+        """
+        Override ``Counter.__setitem__()`` to invalidate the cached N
+        """
+        self._N = None
+        super(FreqDist, self).__setitem__(key, val)
+
+    def __delitem__(self, key):
+        """
+        Override ``Counter.__delitem__()`` to invalidate the cached N
+        """
+        self._N = None
+        super(FreqDist, self).__delitem__(key)
+
+    def update(self, *args, **kwargs):
+        """
+        Override ``Counter.update()`` to invalidate the cached N
+        """
+        self._N = None
+        super(FreqDist, self).update(*args, **kwargs)
+
+    def setdefault(self, key, val):
+        """
+        Override ``Counter.setdefault()`` to invalidate the cached N
+        """
+        self._N = None
+        super(FreqDist, self).setdefault(key, val)
 
     def B(self):
         """
@@ -192,9 +226,10 @@ def freq(self, sample):
         :type sample: any
         :rtype: float
         """
-        if self.N() == 0:
+        n = self.N()
+        if n == 0:
             return 0
-        return self[sample] / self.N()
+        return self[sample] / n
 
     def max(self):
         """
@@ -1749,6 +1784,7 @@ def __init__(self, cond_samples=None):
         :type cond_samples: Sequence of (condition, sample) tuples
         """
         defaultdict.__init__(self, FreqDist)
+
         if cond_samples:
             for (cond, sample) in cond_samples:
                 self[cond][sample] += 1
diff --git a/nltk/tag/tnt.py b/nltk/tag/tnt.py
index 0347b8df23..63db23a3a0 100755
--- a/nltk/tag/tnt.py
+++ b/nltk/tag/tnt.py
@@ -357,30 +357,24 @@ def _tagword(self, sent, current_states):
         # if word is known
         # compute the set of possible tags
         # and their associated log probabilities
-        if word in self._wd.conditions():
+        if word in self._wd:
             self.known += 1
 
             for (history, curr_sent_logprob) in current_states:
                 logprobs = []
 
                 for t in self._wd[word].keys():
-                    p_uni = self._uni.freq((t,C))
-                    p_bi = self._bi[history[-1]].freq((t,C))
-                    p_tri = self._tri[tuple(history[-2:])].freq((t,C))
-                    p_wd = self._wd[word][t] / self._uni[(t,C)]
+                    tC = (t,C)
+                    p_uni = self._uni.freq(tC)
+                    p_bi = self._bi[history[-1]].freq(tC)
+                    p_tri = self._tri[tuple(history[-2:])].freq(tC)
+                    p_wd = self._wd[word][t] / self._uni[tC]
                     p = self._l1 *p_uni + self._l2 *p_bi + self._l3 *p_tri
                     p2 = log(p, 2) + log(p_wd, 2)
 
-                    logprobs.append(((t,C), p2))
-
-
-                # compute the result of appending each tag to this history
-                for (tag, logprob) in logprobs:
-                    new_states.append((history + [tag],
-                                       curr_sent_logprob + logprob))
-
-
-
+                    # compute the result of appending each tag to this history
+                    new_states.append((history + [tC],
+                                       curr_sent_logprob + p2))
 
         # otherwise a new word, set of possible tags is unknown
         else:
@@ -398,7 +392,7 @@ def _tagword(self, sent, current_states):
                 tag = ('Unk',C)
 
             # otherwise apply the unknown word tagger
-            else :
+            else:
                 [(_w, t)] = list(self._unk.tag([word]))
                 tag = (t,C)
 
@@ -407,8 +401,6 @@ def _tagword(self, sent, current_states):
 
             new_states = current_states
 
-
-
         # now have computed a set of possible new_states
 
         # sort states by log prob
@@ -420,7 +412,6 @@ def _tagword(self, sent, current_states):
         if len(new_states) > self._N:
             new_states = new_states[:self._N]
 
-
         # compute the tags for the rest of the sentence
         # return the best list of tags for the sentence
         return self._tagword(sent, new_states)