diff --git a/src/fast_align.cc b/src/fast_align.cc
index 3b4f827..c45893a 100644
--- a/src/fast_align.cc
+++ b/src/fast_align.cc
@@ -58,6 +58,7 @@ void ParseLine(const string& line,
 
 string input;
 string conditional_probability_filename = "";
+string existing_probability_filename = "";
 int is_reverse = 0;
 int ITERATIONS = 5;
 int favor_diagonal = 0;
@@ -79,13 +80,14 @@ struct option options[] = {
     {"alpha",             required_argument, 0,                  'a'},
     {"no_null_word",      no_argument,       &no_null_word,      1  },
     {"conditional_probabilities", required_argument, 0,          'c'},
+    {"existing_probabilities", required_argument, 0,             'e'},
     {0,0,0,0}
 };
 
 bool InitCommandLine(int argc, char** argv) {
   while (1) {
     int oi;
-    int c = getopt_long(argc, argv, "i:rI:dp:T:ova:Nc:", options, &oi);
+    int c = getopt_long(argc, argv, "i:rI:dp:T:ova:Nc:e:", options, &oi);
     if (c == -1) break;
     switch(c) {
       case 'i': input = optarg; break;
@@ -99,6 +101,7 @@ bool InitCommandLine(int argc, char** argv) {
       case 'a': alpha = atof(optarg); break;
       case 'N': no_null_word = 1; break;
       case 'c': conditional_probability_filename = optarg; break;
+      case 'e': existing_probability_filename = optarg; break;
       default: return false;
     }
   }
@@ -116,6 +119,7 @@ int main(int argc, char** argv) {
          << "  -o: [USE] Optimize how close to the diagonal alignment points should be\n"
          << "  -r: Run alignment in reverse (condition on target and predict source)\n"
          << "  -c: Output conditional probability table\n"
+         << "  -e: Start with existing conditional probability table\n"
          << " Advanced options:\n"
          << "  -I: number of iterations in EM training (default = 5)\n"
          << "  -p: p_null parameter (default = 0.08)\n"
@@ -132,12 +136,19 @@ int main(int argc, char** argv) {
   double prob_align_not_null = 1.0 - prob_align_null;
   const unsigned kNULL = d.Convert("<eps>");
   TTable s2t, t2s;
+  if (!existing_probability_filename.empty()) {
+    bool success = s2t.ImportFromFile(existing_probability_filename.c_str(), '\t', d);
+    if (!success) {
+      cerr << "Can't read " << existing_probability_filename << endl;
+      return 1;
+    }
+  }
   unordered_map<pair<short, short>, unsigned, PairHash> size_counts;
   double tot_len_ratio = 0;
   double mean_srclen_multiplier = 0;
   vector<double> probs;
-  for (int iter = 0; iter < ITERATIONS; ++iter) {
-    const bool final_iteration = (iter == (ITERATIONS - 1));
+  for (int iter = 0; iter < ITERATIONS || (iter==0 && ITERATIONS==0); ++iter) {
+    const bool final_iteration = (iter >= (ITERATIONS - 1));
     cerr << "ITERATION " << (iter + 1) << (final_iteration ? " (FINAL)" : "") << endl;
     ifstream in(input.c_str());
     if (!in) {
diff --git a/src/ttables.h b/src/ttables.h
index f5f1cb7..f2a2cb9 100644
--- a/src/ttables.h
+++ b/src/ttables.h
@@ -17,6 +17,9 @@
 
 #include <cmath>
 #include <fstream>
+#include <sstream>
+#include <stdlib.h>
+#include <string>
 #include <tr1/unordered_map>
 
 struct Md {
@@ -107,6 +110,35 @@ class TTable {
     }
     file.close();
   }
+  bool ImportFromFile(const char* filename, char delim, Dict& d) {
+    std::ifstream in(filename);
+    if (!in) {
+      return false;
+    } else {
+      std::string line;
+      while(true) {
+	std::getline(in, line);
+	if (!in) break;
+	std::string sourceWord, targetWord, valueString;
+	std::stringstream stream(line);
+	
+	bool success = true;
+	success &= (std::getline(stream, sourceWord, delim) != NULL);
+	success &= (std::getline(stream, targetWord, delim) != NULL);
+	success &= (std::getline(stream, valueString, delim) != NULL);
+	
+	if (success) {
+	  unsigned source = d.Convert(sourceWord);
+	  unsigned target = d.Convert(targetWord);
+	  double value = atof(valueString.c_str());
+	  ttable[source][target] = value;
+	} else {
+	  return false;
+	}
+      }
+    }
+    return true;
+  }
  public:
   Word2Word2Double ttable;
   Word2Word2Double counts;