diff --git a/nw/core/tokenizer.py b/nw/core/tokenizer.py
index d0bc33eb..708a23de 100644
--- a/nw/core/tokenizer.py
+++ b/nw/core/tokenizer.py
@@ -366,6 +366,12 @@ class Tokenizer():
             (QRegularExpression(nwRegEx.FMT_ST), [None, self.FMT_D_B, None, self.FMT_D_E]),
         ]
 
+        # Determine default text alignment
+        if self.isTitle or self.isPart:
+            defAlign = self.A_CENTRE
+        else:
+            defAlign = self.A_NONE
+
         self.theTokens = []
         tmpMarkdown = []
         nLine = 0
@@ -375,7 +381,7 @@ class Tokenizer():
             # Tag lines starting with specific characters
             if len(aLine.strip()) == 0:
                 self.theTokens.append((
-                    self.T_EMPTY, nLine, "", None, self.A_NONE
+                    self.T_EMPTY, nLine, "", None, defAlign
                 ))
                 if self.keepMarkdown:
                     tmpMarkdown.append("\n")
@@ -385,48 +391,48 @@ class Tokenizer():
                 synTag = cLine[:9].lower()
                 if synTag == "synopsis:":
                     self.theTokens.append((
-                        self.T_SYNOPSIS, nLine, cLine[9:].strip(), None, self.A_NONE
+                        self.T_SYNOPSIS, nLine, cLine[9:].strip(), None, defAlign
                     ))
                     if self.doSynopsis and self.keepMarkdown:
                         tmpMarkdown.append("%s\n" % aLine)
                 else:
                     self.theTokens.append((
-                        self.T_COMMENT, nLine, aLine[1:].strip(), None, self.A_NONE
+                        self.T_COMMENT, nLine, aLine[1:].strip(), None, defAlign
                     ))
                     if self.doComments and self.keepMarkdown:
                         tmpMarkdown.append("%s\n" % aLine)
 
             elif aLine[0] == "@":
                 self.theTokens.append((
-                    self.T_KEYWORD, nLine, aLine[1:].strip(), None, self.A_NONE
+                    self.T_KEYWORD, nLine, aLine[1:].strip(), None, defAlign
                 ))
                 if self.doKeywords and self.keepMarkdown:
                     tmpMarkdown.append("%s\n" % aLine)
 
             elif aLine[:2] == "# ":
                 self.theTokens.append((
-                    self.T_HEAD1, nLine, aLine[2:].strip(), None, self.A_NONE
+                    self.T_HEAD1, nLine, aLine[2:].strip(), None, defAlign
                 ))
                 if self.keepMarkdown:
                     tmpMarkdown.append("%s\n" % aLine)
 
             elif aLine[:3] == "## ":
                 self.theTokens.append((
-                    self.T_HEAD2, nLine, aLine[3:].strip(), None, self.A_NONE
+                    self.T_HEAD2, nLine, aLine[3:].strip(), None, defAlign
                 ))
                 if self.keepMarkdown:
                     tmpMarkdown.append("%s\n" % aLine)
 
             elif aLine[:4] == "### ":
                 self.theTokens.append((
-                    self.T_HEAD3, nLine, aLine[4:].strip(), None, self.A_NONE
+                    self.T_HEAD3, nLine, aLine[4:].strip(), None, defAlign
                 ))
                 if self.keepMarkdown:
                     tmpMarkdown.append("%s\n" % aLine)
 
             elif aLine[:5] == "#### ":
                 self.theTokens.append((
-                    self.T_HEAD4, nLine, aLine[5:].strip(), None, self.A_NONE
+                    self.T_HEAD4, nLine, aLine[5:].strip(), None, defAlign
                 ))
                 if self.keepMarkdown:
                     tmpMarkdown.append("%s\n" % aLine)
@@ -453,7 +459,7 @@ class Tokenizer():
                     tagLeft = True
                     aLine = aLine[:-8].rstrip()
 
-                textAlign = self.A_NONE
+                textAlign = defAlign
                 if tagLeft and tagRight:
                     textAlign = self.A_CENTRE
                 elif tagLeft:
@@ -484,7 +490,7 @@ class Tokenizer():
 
         # Always add an empty line at the end
         self.theTokens.append((
-            self.T_EMPTY, nLine, "", None, self.A_NONE
+            self.T_EMPTY, nLine, "", None, defAlign
         ))
         if self.keepMarkdown:
             tmpMarkdown.append("\n")
@@ -496,8 +502,8 @@ class Tokenizer():
         # ===========
         # Some items need a second pass
 
-        pToken = (self.T_EMPTY, 0, "", None, self.A_NONE)
-        nToken = (self.T_EMPTY, 0, "", None, self.A_NONE)
+        pToken = (self.T_EMPTY, 0, "", None, defAlign)
+        nToken = (self.T_EMPTY, 0, "", None, defAlign)
         tCount = len(self.theTokens)
         for n, tToken in enumerate(self.theTokens):
 
@@ -626,27 +632,27 @@ class Tokenizer():
                             tToken[0], tToken[1], tTemp, None, self.A_NONE
                         )
 
-        # For title page and partitions, we need to centre all text.
-        # For partition, we also add a page break before, and for
-        # both types we always add a page break after the content.
-        # We also swap header level 1 with a title type instead.
+        # For title page we use a different title class, and we will set
+        # an automatic page break before (i.e. added if needed). For
+        # partitions we always need a page break before.
         if self.isTitle or self.isPart:
             for n, tToken in enumerate(self.theTokens):
+                aStyle = tToken[4]
+                if n == 0:
+                    if self.isTitle:
+                        aStyle |= self.A_PBB_AUT
+                    else:
+                        aStyle |= self.A_PBB
+
                 if tToken[0] == self.T_HEAD1:
                     if self.isTitle:
-                        aStyle = self.A_PBB_AUT | self.A_CENTRE
                         self.theTokens[n] = (
                             self.T_TITLE, tToken[1], tToken[2], tToken[3], aStyle
                         )
                     else:
-                        aStyle = self.A_PBB | self.A_CENTRE
                         self.theTokens[n] = (
                             tToken[0], tToken[1], tToken[2], tToken[3], aStyle
                         )
-                else:
-                    self.theTokens[n] = (
-                        tToken[0], tToken[1], tToken[2], tToken[3], self.A_CENTRE
-                    )
 
             # Add a page break after the last entry
             n = len(self.theTokens) - 1
diff --git a/sample/content/53b69b83cdafc.nwd b/sample/content/53b69b83cdafc.nwd
index f33a2a98..cca928b9 100644
--- a/sample/content/53b69b83cdafc.nwd
+++ b/sample/content/53b69b83cdafc.nwd
@@ -5,4 +5,6 @@
 
 **By Jane Doh**
 
-It’s also possible to add some text on this page.
\ No newline at end of file
+It’s also possible to add some text on this page. All text on Title and Partition pages are by default centred when the document is exported, but this can be overridden on individual paragraphs.
+
+The text can be left-aligned like this. <<
diff --git a/sample/nwProject.nwx b/sample/nwProject.nwx
index a5e489bc..8512ed42 100644
--- a/sample/nwProject.nwx
+++ b/sample/nwProject.nwx
@@ -1,13 +1,13 @@
 <?xml version='1.0' encoding='utf-8'?>
-<novelWriterXML appVersion="1.4a0" hexVersion="0x010400a0" fileVersion="1.2" timeStamp="2021-06-09 22:06:41">
+<novelWriterXML appVersion="1.4a0" hexVersion="0x010400a0" fileVersion="1.2" timeStamp="2021-06-09 22:47:04">
   <project>
     <name>Sample Project</name>
     <title>Sample Project</title>
     <author>Jane Smith</author>
     <author>Jay Doh</author>
-    <saveCount>1083</saveCount>
-    <autoCount>172</autoCount>
-    <editTime>50790</editTime>
+    <saveCount>1088</saveCount>
+    <autoCount>176</autoCount>
+    <editTime>51413</editTime>
   </project>
   <settings>
     <doBackup>False</doBackup>
@@ -17,8 +17,8 @@
     <autoOutline>True</autoOutline>
     <lastEdited>636b6aa9b697b</lastEdited>
     <lastViewed>636b6aa9b697b</lastViewed>
-    <lastWordCount>1132</lastWordCount>
-    <novelWordCount>756</novelWordCount>
+    <lastWordCount>1164</lastWordCount>
+    <novelWordCount>788</novelWordCount>
     <notesWordCount>376</notesWordCount>
     <autoReplace>
       <entry key="A">B</entry>
@@ -63,10 +63,10 @@
       <status>Started</status>
       <exported>True</exported>
       <layout>TITLE</layout>
-      <charCount>72</charCount>
-      <wordCount>15</wordCount>
-      <paraCount>2</paraCount>
-      <cursorPos>78</cursorPos>
+      <charCount>259</charCount>
+      <wordCount>47</wordCount>
+      <paraCount>3</paraCount>
+      <cursorPos>268</cursorPos>
     </item>
     <item handle="974e400180a99" order="1" parent="7031beac91f75">
       <name>Page</name>