diff --git a/src/display/context.rs b/src/display/context.rs
index 74148758e7..b8472e2fb9 100644
--- a/src/display/context.rs
+++ b/src/display/context.rs
@@ -143,7 +143,7 @@ fn matched_lines_from_unchanged(
     for mp in mps {
         let opposite_line = match &mp.kind {
             MatchKind::UnchangedToken { opposite_pos, .. }
-            | MatchKind::NovelLinePart { opposite_pos, .. } => {
+            | MatchKind::UnchangedPartOfNovelItem { opposite_pos, .. } => {
                 if let Some(highest_opposite_side) = highest_opposite_line {
                     opposite_pos
                         .iter()
@@ -344,7 +344,7 @@ pub(crate) fn opposite_positions(
                     opposite_lines.insert(opposite_span.line);
                 }
             }
-            MatchKind::NovelLinePart {
+            MatchKind::UnchangedPartOfNovelItem {
                 opposite_pos,
                 self_pos,
                 ..
@@ -837,7 +837,7 @@ mod tests {
     fn test_all_lines() {
         let mps = [
             MatchedPos {
-                kind: MatchKind::NovelLinePart {
+                kind: MatchKind::UnchangedPartOfNovelItem {
                     highlight: TokenKind::Delimiter,
                     self_pos: SingleLineSpan {
                         line: 0.into(),
diff --git a/src/display/json.rs b/src/display/json.rs
index 2dc867dee1..f0a10498c8 100644
--- a/src/display/json.rs
+++ b/src/display/json.rs
@@ -255,7 +255,7 @@ impl Highlight {
             MatchKind::UnchangedToken { highlight, .. } => highlight,
             MatchKind::Novel { highlight, .. } => highlight,
             MatchKind::NovelWord { highlight, .. } => highlight,
-            MatchKind::NovelLinePart { highlight, .. } => highlight,
+            MatchKind::UnchangedPartOfNovelItem { highlight, .. } => highlight,
         };
 
         match highlight {
diff --git a/src/display/style.rs b/src/display/style.rs
index 7beb5d3fef..8f31db145d 100644
--- a/src/display/style.rs
+++ b/src/display/style.rs
@@ -393,7 +393,7 @@ pub(crate) fn color_positions(
                     style = style.italic();
                 }
             }
-            MatchKind::NovelLinePart { highlight, .. } => {
+            MatchKind::UnchangedPartOfNovelItem { highlight, .. } => {
                 style = novel_style(style, side, background);
                 if syntax_highlight && matches!(highlight, TokenKind::Atom(AtomKind::Comment)) {
                     style = style.italic();
diff --git a/src/line_parser.rs b/src/line_parser.rs
index f2ed5cd335..b546294a7e 100644
--- a/src/line_parser.rs
+++ b/src/line_parser.rs
@@ -192,7 +192,7 @@ pub(crate) fn change_positions(lhs_src: &str, rhs_src: &str) -> Vec<MatchedPos>
                                     rhs_lp.from_region(rhs_offset, rhs_offset + rhs_word.len());
 
                                 mps.push(MatchedPos {
-                                    kind: MatchKind::NovelLinePart {
+                                    kind: MatchKind::UnchangedPartOfNovelItem {
                                         highlight: TokenKind::Atom(AtomKind::Normal),
                                         self_pos: lhs_pos[0],
                                         opposite_pos: rhs_pos,
diff --git a/src/parse/syntax.rs b/src/parse/syntax.rs
index 1f08333457..ab2cafcc6a 100644
--- a/src/parse/syntax.rs
+++ b/src/parse/syntax.rs
@@ -598,15 +598,33 @@ pub(crate) enum MatchKind {
     },
     /// A novel token in an AST diff.
     Novel { highlight: TokenKind },
-    /// When doing a textual line-based diff, the part of a novel line
-    /// that exists on both sides.
-    NovelLinePart {
+    /// When we have a novel item, we often want to highlight novel
+    /// words more prominently. UnchangedPartOfNovelItem represents
+    /// the parts that don't get this special highlighting.
+    ///
+    /// For example, line-based diffs we want to highlight `a` and `b`
+    /// differently to `foo` here.
+    ///
+    /// foo a
+    /// foo b
+    ///
+    /// Whereas for syntactic diffs, we want to do the same thing for
+    /// strings and comments.
+    ///
+    /// "foo a"
+    /// "foo b"
+    ///
+    /// The whole string is a distinct value, but the `a` and `b` are
+    /// the most interesting parts.
+    UnchangedPartOfNovelItem {
         highlight: TokenKind,
         self_pos: SingleLineSpan,
         opposite_pos: Vec<SingleLineSpan>,
     },
-    /// When doing a textual line-based diff, the part of a novel line
-    /// that is actually novel. E.g. the newly added word on a line.
+    /// The novel part of the novel item. For line-based diffs, this
+    /// is the words that are unique to this line.
+    ///
+    /// See the discussion in `UnchangedPartOfNovelItem`.
     NovelWord { highlight: TokenKind },
     /// A syntactic token that was ignored by the AST diff (e.g. when
     /// ignoring comments for diffing).
@@ -617,7 +635,9 @@ impl MatchKind {
     pub(crate) fn is_novel(&self) -> bool {
         matches!(
             self,
-            MatchKind::Novel { .. } | MatchKind::NovelWord { .. } | MatchKind::NovelLinePart { .. }
+            MatchKind::Novel { .. }
+                | MatchKind::NovelWord { .. }
+                | MatchKind::UnchangedPartOfNovelItem { .. }
         )
     }
 }
@@ -700,7 +720,7 @@ fn split_atom_words(
                 );
 
                 mps.push(MatchedPos {
-                    kind: MatchKind::NovelLinePart {
+                    kind: MatchKind::UnchangedPartOfNovelItem {
                         highlight: TokenKind::Atom(kind),
                         self_pos: word_pos,
                         opposite_pos: opposite_word_pos,
@@ -1184,7 +1204,7 @@ mod tests {
             res,
             vec![
                 MatchedPos {
-                    kind: MatchKind::NovelLinePart {
+                    kind: MatchKind::UnchangedPartOfNovelItem {
                         highlight: TokenKind::Atom(AtomKind::Comment),
                         self_pos: SingleLineSpan {
                             line: 0.into(),
@@ -1204,7 +1224,7 @@ mod tests {
                     }
                 },
                 MatchedPos {
-                    kind: MatchKind::NovelLinePart {
+                    kind: MatchKind::UnchangedPartOfNovelItem {
                         highlight: TokenKind::Atom(AtomKind::Comment),
                         self_pos: SingleLineSpan {
                             line: 0.into(),
@@ -1224,7 +1244,7 @@ mod tests {
                     }
                 },
                 MatchedPos {
-                    kind: MatchKind::NovelLinePart {
+                    kind: MatchKind::UnchangedPartOfNovelItem {
                         highlight: TokenKind::Atom(AtomKind::Comment),
                         self_pos: SingleLineSpan {
                             line: 0.into(),
@@ -1244,7 +1264,7 @@ mod tests {
                     }
                 },
                 MatchedPos {
-                    kind: MatchKind::NovelLinePart {
+                    kind: MatchKind::UnchangedPartOfNovelItem {
                         highlight: TokenKind::Atom(AtomKind::Comment),
                         self_pos: SingleLineSpan {
                             line: 0.into(),
@@ -1264,7 +1284,7 @@ mod tests {
                     }
                 },
                 MatchedPos {
-                    kind: MatchKind::NovelLinePart {
+                    kind: MatchKind::UnchangedPartOfNovelItem {
                         highlight: TokenKind::Atom(AtomKind::Comment),
                         self_pos: SingleLineSpan {
                             line: 0.into(),