diff --git a/compare50/_data.py b/compare50/_data.py
index c9b17d7..2d0dbd9 100644
--- a/compare50/_data.py
+++ b/compare50/_data.py
@@ -1,12 +1,12 @@
 import abc
 from collections.abc import Mapping, Sequence
-import os
 import pathlib
 import numbers
 
 import attr
 import pygments
 import pygments.lexers
+from .lexers import WordLexer
 
 
 __all__ = ["Pass", "Comparator", "File", "Submission",
@@ -191,14 +191,18 @@ def lexer(self):
 
         # get lexer for this file type
         try:
-            lexer = pygments.lexers.get_lexer_for_filename(self.name.name)
+            if ext == ".txt":
+                lexer = WordLexer()
+            else:
+                lexer = pygments.lexers.get_lexer_for_filename(self.name.name)
+    
             self._lexer_cache[ext] = lexer
             return lexer
         except pygments.util.ClassNotFound:
             try:
                 return pygments.lexers.guess_lexer(self.read())
             except pygments.util.ClassNotFound:
-                return pygments.lexers.special.TextLexer()
+                return WordLexer()
 
     @classmethod
     def get(cls, id):
diff --git a/compare50/lexers.py b/compare50/lexers.py
new file mode 100644
index 0000000..fadf605
--- /dev/null
+++ b/compare50/lexers.py
@@ -0,0 +1,16 @@
+from pygments.lexer import RegexLexer
+from pygments.token import Text, Name
+
+class WordLexer(RegexLexer):
+    """Custom compare50 lexer that creates a token based on each 'word'."""
+    name = "WordLexer"
+    aliases = ["word"]
+    filenames = ["*.txt"]
+
+    tokens = {
+        "root": [
+            (r"\s+", Text),           # whitespace
+            (r"\w+", Name),           # word (alphanumeric)
+            (r"\W", Text),            # punctuation or other
+        ]
+    }