diff --git a/spec/text-editor-registry-spec.js b/spec/text-editor-registry-spec.js
index 8ecb40c5b..b4c64db10 100644
--- a/spec/text-editor-registry-spec.js
+++ b/spec/text-editor-registry-spec.js
@@ -75,7 +75,7 @@ describe('TextEditorRegistry', function () {
     })
 
     it('updates the editor\'s grammar when a more appropriate grammar is added for its path', async function () {
-      expect(editor.getGrammar()).toBe(null)
+      expect(editor.getGrammar().name).toBe('Null Grammar')
 
       editor.getBuffer().setPath('test.js')
       registry.maintainGrammar(editor)
@@ -473,13 +473,13 @@ describe('TextEditorRegistry', function () {
       const editor2Copy = TextEditor.deserialize(editor2.serialize(), atom)
       const registryCopy = TextEditorRegistry.deserialize(registry.serialize(), atom)
 
-      expect(editorCopy.getGrammar()).toBe(null)
-      expect(editor2Copy.getGrammar()).toBe(null)
+      expect(editorCopy.getGrammar().name).toBe('Null Grammar')
+      expect(editor2Copy.getGrammar().name).toBe('Null Grammar')
 
       registryCopy.maintainGrammar(editorCopy)
       registryCopy.maintainGrammar(editor2Copy)
       expect(editorCopy.getGrammar().name).toBe('C')
-      expect(editor2Copy.getGrammar()).toBe(null)
+      expect(editor2Copy.getGrammar().name).toBe('Null Grammar')
 
       await atom.packages.activatePackage('language-javascript')
       expect(editorCopy.getGrammar().name).toBe('C')
diff --git a/src/null-grammar.js b/src/null-grammar.js
new file mode 100644
index 000000000..33f7a5dc4
--- /dev/null
+++ b/src/null-grammar.js
@@ -0,0 +1,4 @@
+module.exports = Object.freeze({
+  name: 'Null Grammar',
+  scopeName: 'text.plain'
+})
diff --git a/src/tokenized-buffer.coffee b/src/tokenized-buffer.coffee
index 90666b783..aa5789016 100644
--- a/src/tokenized-buffer.coffee
+++ b/src/tokenized-buffer.coffee
@@ -8,10 +8,11 @@ TokenIterator = require './token-iterator'
 Token = require './token'
 ScopeDescriptor = require './scope-descriptor'
 TokenizedBufferIterator = require './tokenized-buffer-iterator'
+NullGrammar = require './null-grammar'
 
 module.exports =
 class TokenizedBuffer extends Model
-  grammar: null
+  grammar: NullGrammar
   buffer: null
   tabLength: null
   tokenizedLines: null
@@ -127,7 +128,7 @@ class TokenizedBuffer extends Model
 
   tokenizeNextChunk: ->
     # Short circuit null grammar which can just use the placeholder tokens
-    if (not @grammar? or @grammar.name is 'Null grammar') and @firstInvalidRow()?
+    if (@grammar.name is 'Null Grammar') and @firstInvalidRow()?
       @invalidRows = []
       @markTokenizationComplete()
       return
@@ -203,7 +204,7 @@ class TokenizedBuffer extends Model
 
     @updateInvalidRows(start, end, delta)
     previousEndStack = @stackForRow(end) # used in spill detection below
-    if @largeFileMode or not @grammar?
+    if @largeFileMode or @grammar is NullGrammar
       newTokenizedLines = @buildPlaceholderTokenizedLinesForRows(start, end + delta)
     else
       newTokenizedLines = @buildTokenizedLinesForRows(start, end + delta, @stackForRow(start - 1), @openScopesForRow(start))
@@ -279,7 +280,7 @@ class TokenizedBuffer extends Model
     @buildPlaceholderTokenizedLineForRow(row) for row in [startRow..endRow] by 1
 
   buildPlaceholderTokenizedLineForRow: (row) ->
-    if @grammar?
+    if @grammar isnt NullGrammar
       openScopes = [@grammar.startIdForScope(@grammar.scopeName)]
     else
       openScopes = []