diff --git a/js/jquery.autocomplete.js b/js/jquery.autocomplete.js
index 860a2f647..effd167a5 100644
--- a/js/jquery.autocomplete.js
+++ b/js/jquery.autocomplete.js
@@ -473,8 +473,14 @@
 
 		function matchSubset(s, sub) {
 			if (!options.matchCase)
-				s = s.toLowerCase().normalize('NFD').replace(/[\u0300-\u036f]/g, "");
-            	sub = sub.toLowerCase().normalize('NFD').replace(/[\u0300-\u036f]/g, "");
+				if (typeof s.normalize === 'function'){
+					s = s.toLowerCase().normalize('NFD').replace(/[\u0300-\u036f]/g, "");
+	            	sub = sub.toLowerCase().normalize('NFD').replace(/[\u0300-\u036f]/g, "");					
+				}
+				else {
+					s = ie_normalize(s.toLowerCase());
+	            	sub = ie_normalize(sub.toLowerCase());										
+				}
 			var i = s.indexOf(sub);
 			if (options.matchContains === "word"){
 				i = s.toLowerCase().search("\\b" + sub.toLowerCase());
@@ -482,7 +488,24 @@
 			if (i === -1) return false;
 			return i === 0 || options.matchContains;
 		};
-
+		
+		function ie_normalize(s)
+		{
+			//Cheap replacement for normalize on IE. Works only on a (small) subset of what's possible in unicode
+		    var r = s.toLowerCase();
+		    r = r.replace(new RegExp(/[àáâãäå]/g),"a");
+		    r = r.replace(new RegExp(/æ/g),"ae");
+		    r = r.replace(new RegExp(/ç/g),"c");
+		    r = r.replace(new RegExp(/[èéêë]/g),"e");
+		    r = r.replace(new RegExp(/[ìíîï]/g),"i");
+		    r = r.replace(new RegExp(/ñ/g),"n");                
+		    r = r.replace(new RegExp(/[òóôõö]/g),"o");
+		    r = r.replace(new RegExp(/œ/g),"oe");
+		    r = r.replace(new RegExp(/[ùúûü]/g),"u");
+		    r = r.replace(new RegExp(/[ýÿ]/g),"y");
+		    return r;			
+		}
+		
 		function add(q, value) {
 			if (length > options.cacheLength){
 				flush();