diff --git a/gensim/models/ldamodel.py b/gensim/models/ldamodel.py
index a1f469aaae..97f3921d05 100755
--- a/gensim/models/ldamodel.py
+++ b/gensim/models/ldamodel.py
@@ -800,15 +800,19 @@ def show_topics(self, num_topics=10, num_words=10, log=False, formatted=True):
             chosen_topics = sorted_topics[:num_topics // 2] + sorted_topics[-num_topics // 2:]
 
         shown = []
+
+        topic = self.state.get_lambda()
         for i in chosen_topics:
+            topic_ = topic[i]
+            topic_ = topic_ / topic_.sum()  # normalize to probability distribution
+            bestn = matutils.argsort(topic_, num_words, reverse=True)
+            topic_ = [(self.id2word[id], topic_[id]) for id in bestn]
             if formatted:
-                topic = self.print_topic(i, topn=num_words)
-            else:
-                topic = self.show_topic(i, topn=num_words)
+                topic_ = ' + '.join(['%.3f*"%s"' % (v, k) for k, v in topic_])
 
-            shown.append((i, topic))
+            shown.append((i, topic_))
             if log:
-                logger.info("topic #%i (%.3f): %s", i, self.alpha[i], topic)
+                logger.info("topic #%i (%.3f): %s", i, self.alpha[i], topic_)
 
         return shown