diff --git a/fonts/SegoeUI-Emoji.ttf b/fonts/SegoeUI-Emoji.ttf
new file mode 100644
index 000000000..77497bf76
Binary files /dev/null and b/fonts/SegoeUI-Emoji.ttf differ
diff --git a/src/layout/mod.rs b/src/layout/mod.rs
index 8a68a6a5d..3cfb872fd 100644
--- a/src/layout/mod.rs
+++ b/src/layout/mod.rs
@@ -142,8 +142,6 @@ pub enum Command<'a> {
 
     /// Start a new line.
     BreakLine,
-    /// Start a new paragraph.
-    BreakParagraph,
     /// Start a new page, which will be part of the finished layout even if it
     /// stays empty (since the page break is a _hard_ space break).
     BreakPage,
diff --git a/src/layout/tree.rs b/src/layout/tree.rs
index d20fc666c..3abdc9342 100644
--- a/src/layout/tree.rs
+++ b/src/layout/tree.rs
@@ -43,26 +43,43 @@ impl<'a> TreeLayouter<'a> {
         }
     }
 
-    async fn layout_tree(&mut self, tree: &SyntaxTree) {
-        for node in tree {
-            self.layout_node(node).await;
-        }
-    }
-
     fn finish(self) -> Pass<MultiLayout> {
         Pass::new(self.layouter.finish(), self.feedback)
     }
 
+    fn layout_tree<'t>(&'t mut self, tree: &'t SyntaxTree) -> DynFuture<'t, ()> {
+        Box::pin(async move {
+            for node in tree {
+                self.layout_node(node).await;
+            }
+        })
+    }
+
     async fn layout_node(&mut self, node: &Spanned<SyntaxNode>) {
-        let decorate = |this: &mut TreeLayouter, deco| {
+        let decorate = |this: &mut Self, deco| {
             this.feedback.decorations.push(Spanned::new(deco, node.span));
         };
 
         match &node.v {
-            SyntaxNode::Space => self.layout_space(),
-            SyntaxNode::Parbreak => self.layout_paragraph(),
+            SyntaxNode::Spacing => {
+                self.layouter.add_primary_spacing(
+                    self.style.text.word_spacing(),
+                    SpacingKind::WORD,
+                );
+            },
+
             SyntaxNode::Linebreak => self.layouter.finish_line(),
 
+            SyntaxNode::ToggleItalic => {
+                self.style.text.italic = !self.style.text.italic;
+                decorate(self, Decoration::Italic);
+            }
+
+            SyntaxNode::ToggleBolder => {
+                self.style.text.bolder = !self.style.text.bolder;
+                decorate(self, Decoration::Bold);
+            }
+
             SyntaxNode::Text(text) => {
                 if self.style.text.italic {
                     decorate(self, Decoration::Italic);
@@ -75,16 +92,6 @@ impl<'a> TreeLayouter<'a> {
                 self.layout_text(text).await;
             }
 
-            SyntaxNode::ToggleItalic => {
-                self.style.text.italic = !self.style.text.italic;
-                decorate(self, Decoration::Italic);
-            }
-
-            SyntaxNode::ToggleBolder => {
-                self.style.text.bolder = !self.style.text.bolder;
-                decorate(self, Decoration::Bold);
-            }
-
             SyntaxNode::Raw(lines) => {
                 // TODO: Make this more efficient.
                 let fallback = self.style.text.fallback.clone();
@@ -109,14 +116,25 @@ impl<'a> TreeLayouter<'a> {
                 self.style.text.fallback = fallback;
             }
 
+            SyntaxNode::Par(par) => self.layout_par(par).await,
+
             SyntaxNode::Dyn(dynamic) => {
                 self.layout_dyn(Spanned::new(dynamic.as_ref(), node.span)).await;
             }
         }
     }
 
+    async fn layout_par(&mut self, par: &SyntaxTree) {
+        self.layouter.add_secondary_spacing(
+            self.style.text.paragraph_spacing(),
+            SpacingKind::PARAGRAPH,
+        );
+
+        self.layout_tree(par).await;
+    }
+
     async fn layout_dyn(&mut self, dynamic: Spanned<&dyn DynamicNode>) {
-        // Execute the tree's command-generating layout function.
+        // Execute the dynamic node's command-generating layout function.
         let layouted = dynamic.v.layout(LayoutContext {
             style: &self.style,
             spaces: self.layouter.remaining(),
@@ -131,11 +149,21 @@ impl<'a> TreeLayouter<'a> {
         }
     }
 
-    fn execute_command<'r>(
-        &'r mut self,
-        command: Command<'r>,
-        tree_span: Span,
-    ) -> DynFuture<'r, ()> { Box::pin(async move {
+    async fn layout_text(&mut self, text: &str) {
+        self.layouter.add(
+            layout_text(
+                text,
+                TextContext {
+                    loader: &self.ctx.loader,
+                    style: &self.style.text,
+                    dir: self.ctx.axes.primary,
+                    align: self.ctx.align,
+                }
+            ).await
+        );
+    }
+
+    async fn execute_command(&mut self, command: Command<'_>, span: Span) {
         use Command::*;
 
         match command {
@@ -149,13 +177,12 @@ impl<'a> TreeLayouter<'a> {
             }
 
             BreakLine => self.layouter.finish_line(),
-            BreakParagraph => self.layout_paragraph(),
             BreakPage => {
                 if self.ctx.root {
                     self.layouter.finish_space(true)
                 } else {
                     error!(
-                        @self.feedback, tree_span,
+                        @self.feedback, span,
                         "page break cannot only be issued from root context",
                     );
                 }
@@ -183,7 +210,7 @@ impl<'a> TreeLayouter<'a> {
                     ], true);
                 } else {
                     error!(
-                        @self.feedback, tree_span,
+                        @self.feedback, span,
                         "page style cannot only be changed from root context",
                     );
                 }
@@ -195,33 +222,5 @@ impl<'a> TreeLayouter<'a> {
                 self.ctx.axes = axes;
             }
         }
-    }) }
-
-    async fn layout_text(&mut self, text: &str) {
-        self.layouter.add(
-            layout_text(
-                text,
-                TextContext {
-                    loader: &self.ctx.loader,
-                    style: &self.style.text,
-                    dir: self.ctx.axes.primary,
-                    align: self.ctx.align,
-                }
-            ).await
-        );
-    }
-
-    fn layout_space(&mut self) {
-        self.layouter.add_primary_spacing(
-            self.style.text.word_spacing(),
-            SpacingKind::WORD,
-        );
-    }
-
-    fn layout_paragraph(&mut self) {
-        self.layouter.add_secondary_spacing(
-            self.style.text.paragraph_spacing(),
-            SpacingKind::PARAGRAPH,
-        );
     }
 }
diff --git a/src/library/mod.rs b/src/library/mod.rs
index 0f61b901c..3b09a7685 100644
--- a/src/library/mod.rs
+++ b/src/library/mod.rs
@@ -22,7 +22,6 @@ pub fn std() -> Scope {
     std.add::<PageFunc>("page");
     std.add::<AlignFunc>("align");
     std.add::<BoxFunc>("box");
-    std.add::<ParBreakFunc>("parbreak");
     std.add::<PageBreakFunc>("pagebreak");
     std.add_with_meta::<SpacingFunc>("h", Horizontal);
     std.add_with_meta::<SpacingFunc>("v", Vertical);
diff --git a/src/library/page.rs b/src/library/page.rs
index 1cc76d7fa..b13f8a1ed 100644
--- a/src/library/page.rs
+++ b/src/library/page.rs
@@ -57,3 +57,12 @@ function! {
         vec![SetPageStyle(style)]
     }
 }
+
+function! {
+    /// `pagebreak`: Ends the current page.
+    #[derive(Debug, Default, Clone, PartialEq)]
+    pub struct PageBreakFunc;
+
+    parse(default)
+    layout(self, ctx, f) { vec![BreakPage] }
+}
diff --git a/src/library/spacing.rs b/src/library/spacing.rs
index 545f39102..6503556f0 100644
--- a/src/library/spacing.rs
+++ b/src/library/spacing.rs
@@ -2,26 +2,6 @@ use crate::layout::SpacingKind;
 use crate::length::ScaleLength;
 use super::*;
 
-function! {
-    /// `parbreak`: Ends the current paragraph.
-    ///
-    /// This has the same effect as two subsequent newlines.
-    #[derive(Debug, Default, Clone, PartialEq)]
-    pub struct ParBreakFunc;
-
-    parse(default)
-    layout(self, ctx, f) { vec![BreakParagraph] }
-}
-
-function! {
-    /// `pagebreak`: Ends the current page.
-    #[derive(Debug, Default, Clone, PartialEq)]
-    pub struct PageBreakFunc;
-
-    parse(default)
-    layout(self, ctx, f) { vec![BreakPage] }
-}
-
 function! {
     /// `h` and `v`: Add horizontal or vertical spacing.
     #[derive(Debug, Clone, PartialEq)]
diff --git a/src/style.rs b/src/style.rs
index 803899197..2a8b6d871 100644
--- a/src/style.rs
+++ b/src/style.rs
@@ -75,8 +75,10 @@ impl Default for TextStyle {
                     "monospace" => ["source code pro", "noto sans mono"],
                     "math" => ["latin modern math", "serif"],
                 },
-                base: ["source sans pro", "noto sans",
-                       "noto emoji", "latin modern math"],
+                base: [
+                    "source sans pro", "noto sans", "segoe ui emoji",
+                    "noto emoji", "latin modern math",
+                ],
             },
             variant: FontVariant {
                 style: FontStyle::Normal,
diff --git a/src/syntax/parsing.rs b/src/syntax/parsing.rs
index 5141e455c..3c8020740 100644
--- a/src/syntax/parsing.rs
+++ b/src/syntax/parsing.rs
@@ -96,6 +96,7 @@ pub struct ParseState {
 /// function's body.
 pub fn parse(src: &str, offset: Pos, state: &ParseState) -> Pass<SyntaxTree> {
     let mut tree = SyntaxTree::new();
+    let mut par = SyntaxTree::new();
     let mut feedback = Feedback::new();
 
     for token in Tokens::new(src, offset, TokenMode::Body) {
@@ -103,10 +104,16 @@ pub fn parse(src: &str, offset: Pos, state: &ParseState) -> Pass<SyntaxTree> {
         let node = match token.v {
             // Starting from two newlines counts as a paragraph break, a single
             // newline does not.
-            Token::Space(newlines) => if newlines >= 2 {
-                SyntaxNode::Parbreak
+            Token::Space(newlines) => if newlines < 2 {
+                SyntaxNode::Spacing
             } else {
-                SyntaxNode::Space
+                // End the current paragraph if it is not empty.
+                if let (Some(first), Some(last)) = (par.first(), par.last()) {
+                    let span = Span::merge(first.span, last.span);
+                    let node = SyntaxNode::Par(std::mem::take(&mut par));
+                    tree.push(Spanned::new(node, span));
+                }
+                continue;
             }
 
             Token::Function { header, body, terminated } => {
@@ -136,6 +143,12 @@ pub fn parse(src: &str, offset: Pos, state: &ParseState) -> Pass<SyntaxTree> {
             }
         };
 
+        par.push(Spanned::new(node, span));
+    }
+
+    if let (Some(first), Some(last)) = (par.first(), par.last()) {
+        let span = Span::merge(first.span, last.span);
+        let node = SyntaxNode::Par(par);
         tree.push(Spanned::new(node, span));
     }
 
@@ -671,8 +684,7 @@ mod tests {
     use Decoration::*;
     use Expr::{Bool, Length as Len, Number as Num};
     use SyntaxNode::{
-        Space as S, Parbreak, Linebreak, ToggleItalic as Italic,
-        ToggleBolder as Bold,
+        Spacing as S, Linebreak, ToggleItalic as Italic, ToggleBolder as Bold,
     };
 
     /// Test whether the given string parses into
@@ -714,10 +726,10 @@ mod tests {
         };
     }
 
-    /// Shorthand for `p!("[val: ...]" => func!("val", ...))`.
+    /// Shorthand for `p!("[val: ...]" => par![func!("val", ...)])`.
     macro_rules! pval {
         ($header:expr => $($tts:tt)*) => {
-            p!(concat!("[val: ", $header, "]") => [func!("val": $($tts)*)]);
+            p!(concat!("[val: ", $header, "]") => [par![func!("val": $($tts)*)]]);
         }
     }
 
@@ -725,7 +737,6 @@ mod tests {
     fn Str(text: &str) -> Expr { Expr::Str(text.to_string()) }
     fn Color(r: u8, g: u8, b: u8, a: u8) -> Expr { Expr::Color(RgbaColor::new(r, g, b, a)) }
     fn ColorStr(color: &str) -> Expr { Expr::Color(RgbaColor::from_str(color).expect("invalid test color")) }
-    fn ColorHealed() -> Expr { Expr::Color(RgbaColor::new_healed(0, 0, 0, 255)) }
     fn Neg(e1: Expr) -> Expr { Expr::Neg(Box::new(Z(e1))) }
     fn Add(e1: Expr, e2: Expr) -> Expr { Expr::Add(Box::new(Z(e1)), Box::new(Z(e2))) }
     fn Sub(e1: Expr, e2: Expr) -> Expr { Expr::Sub(Box::new(Z(e1)), Box::new(Z(e2))) }
@@ -764,6 +775,12 @@ mod tests {
         };
     }
 
+    macro_rules! par {
+        ($($tts:tt)*) => {
+            SyntaxNode::Par(span_vec![$($tts)*].0)
+        };
+    }
+
     macro_rules! func {
         ($name:tt
             $(: ($($pos:tt)*) $(, { $($key:tt => $value:expr),* })? )?
@@ -837,32 +854,32 @@ mod tests {
     }
 
     #[test]
-    fn parse_basic_SyntaxNodes() {
-        // Basic SyntaxNodes.
+    fn parse_basic_nodes() {
+        // Basic nodes.
         p!(""                     => []);
-        p!("hi"                   => [T("hi")]);
-        p!("*hi"                  => [Bold, T("hi")]);
-        p!("hi_"                  => [T("hi"), Italic]);
-        p!("hi you"               => [T("hi"), S, T("you")]);
-        p!("hi// you\nw"          => [T("hi"), S, T("w")]);
-        p!("\n\n\nhello"          => [Parbreak, T("hello")]);
-        p!("first//\n//\nsecond"  => [T("first"), S, S, T("second")]);
-        p!("first//\n \nsecond"   => [T("first"), Parbreak, T("second")]);
-        p!("first/*\n \n*/second" => [T("first"), T("second")]);
-        p!(r"a\ b"                => [T("a"), Linebreak, S, T("b")]);
-        p!("💜\n\n 🌍"            => [T("💜"), Parbreak, T("🌍")]);
+        p!("hi"                   => [par![T("hi")]]);
+        p!("*hi"                  => [par![Bold, T("hi")]]);
+        p!("hi_"                  => [par![T("hi"), Italic]]);
+        p!("hi you"               => [par![T("hi"), S, T("you")]]);
+        p!("hi// you\nw"          => [par![T("hi"), S, T("w")]]);
+        p!("\n\n\nhello"          => [par![T("hello")]]);
+        p!("first//\n//\nsecond"  => [par![T("first"), S, S, T("second")]]);
+        p!("first//\n \nsecond"   => [par![T("first")], par![T("second")]]);
+        p!("first/*\n \n*/second" => [par![T("first"), T("second")]]);
+        p!(r"a\ b"                => [par![T("a"), Linebreak, S, T("b")]]);
+        p!("💜\n\n 🌍"           => [par![T("💜")], par![T("🌍")]]);
 
         // Raw markup.
-        p!("`py`"         => [raw!["py"]]);
-        p!("[val][`hi]`]" => [func!("val"; [raw!["hi]"]])]);
-        p!("`hi\nyou"     => [raw!["hi", "you"]], [(1:3, 1:3, "expected backtick")]);
-        p!("`hi\\`du`"    => [raw!["hi`du"]]);
+        p!("`py`"         => [par![raw!["py"]]]);
+        p!("[val][`hi]`]" => [par![func!("val"; [par![raw!["hi]"]]])]]);
+        p!("`hi\nyou"     => [par![raw!["hi", "you"]]], [(1:3, 1:3, "expected backtick")]);
+        p!("`hi\\`du`"    => [par![raw!["hi`du"]]]);
 
         // Spanned SyntaxNodes.
-        p!("Hi"      => [(0:0, 0:2, T("Hi"))]);
-        p!("*Hi*"    => [(0:0, 0:1, Bold), (0:1, 0:3, T("Hi")), (0:3, 0:4, Bold)]);
+        p!("Hi"      => [(0:0, 0:2, par![(0:0, 0:2, T("Hi"))])]);
+        p!("*Hi*"    => [(0:0, 0:4, par![(0:0, 0:1, Bold), (0:1, 0:3, T("Hi")), (0:3, 0:4, Bold)])]);
         p!("🌎\n*/[n]" =>
-            [(0:0, 0:1, T("🌎")), (0:1, 1:0, S), (1:2, 1:5, func!((0:1, 0:2, "n")))],
+            [(0:0, 1:5, par![(0:0, 0:1, T("🌎")), (0:1, 1:0, S), (1:2, 1:5, func!((0:1, 0:2, "n")))])],
             [(1:0, 1:2, "unexpected end of block comment")],
             [(1:3, 1:4, ResolvedFunc)],
         );
@@ -871,52 +888,52 @@ mod tests {
     #[test]
     fn parse_function_names() {
         // No closing bracket.
-        p!("[" => [func!("")], [
+        p!("[" => [par![func!("")]], [
             (0:1, 0:1, "expected function name"),
             (0:1, 0:1, "expected closing bracket")
         ]);
 
         // No name.
-        p!("[]" => [func!("")], [(0:1, 0:1, "expected function name")]);
-        p!("[\"]" => [func!("")], [
+        p!("[]" => [par![func!("")]], [(0:1, 0:1, "expected function name")]);
+        p!("[\"]" => [par![func!("")]], [
             (0:1, 0:3, "expected function name, found string"),
             (0:3, 0:3, "expected closing bracket"),
         ]);
 
         // An unknown name.
         p!("[hi]" =>
-            [func!("hi")],
+            [par![func!("hi")]],
             [(0:1, 0:3, "unknown function")],
             [(0:1, 0:3, UnresolvedFunc)],
         );
 
         // A valid name.
-        p!("[f]"   => [func!("f")], [], [(0:1, 0:2, ResolvedFunc)]);
-        p!("[  f]" => [func!("f")], [], [(0:3, 0:4, ResolvedFunc)]);
+        p!("[f]"   => [par![func!("f")]], [], [(0:1, 0:2, ResolvedFunc)]);
+        p!("[  f]" => [par![func!("f")]], [], [(0:3, 0:4, ResolvedFunc)]);
 
         // An invalid token for a name.
-        p!("[12]"   => [func!("")], [(0:1, 0:3, "expected function name, found number")], []);
-        p!("[🌎]"   => [func!("")], [(0:1, 0:2, "expected function name, found invalid token")], []);
-        p!("[  🌎]" => [func!("")], [(0:3, 0:4, "expected function name, found invalid token")], []);
+        p!("[12]"   => [par![func!("")]], [(0:1, 0:3, "expected function name, found number")], []);
+        p!("[🌎]"   => [par![func!("")]], [(0:1, 0:2, "expected function name, found invalid token")], []);
+        p!("[  🌎]" => [par![func!("")]], [(0:3, 0:4, "expected function name, found invalid token")], []);
     }
 
     #[test]
     fn parse_colon_starting_function_arguments() {
         // Valid.
         p!("[val: true]" =>
-            [func!["val": (Bool(true))]], [],
+            [par![func!["val": (Bool(true))]]], [],
             [(0:1, 0:4, ResolvedFunc)],
         );
 
         // No colon before arg.
-        p!("[val\"s\"]" => [func!("val")], [(0:4, 0:4, "expected colon")]);
+        p!("[val\"s\"]" => [par![func!("val")]], [(0:4, 0:4, "expected colon")]);
 
         // No colon before valid, but wrong token.
-        p!("[val=]" => [func!("val")], [(0:4, 0:4, "expected colon")]);
+        p!("[val=]" => [par![func!("val")]], [(0:4, 0:4, "expected colon")]);
 
         // No colon before invalid tokens, which are ignored.
         p!("[val/🌎:$]" =>
-            [func!("val")],
+            [par![func!("val")]],
             [(0:4, 0:4, "expected colon")],
             [(0:1, 0:4, ResolvedFunc)],
         );
@@ -924,18 +941,18 @@ mod tests {
         // String in invalid header without colon still parsed as string
         // Note: No "expected quote" error because not even the string was
         //       expected.
-        p!("[val/\"]" => [func!("val")], [
+        p!("[val/\"]" => [par![func!("val")]], [
             (0:4, 0:4, "expected colon"),
             (0:7, 0:7, "expected closing bracket"),
         ]);
 
         // Just colon without args.
-        p!("[val:]"         => [func!("val")]);
-        p!("[val:/*12pt*/]" => [func!("val")]);
+        p!("[val:]"         => [par![func!("val")]]);
+        p!("[val:/*12pt*/]" => [par![func!("val")]]);
 
         // Whitespace / comments around colon.
-        p!("[val\n:\ntrue]"      => [func!("val": (Bool(true)))]);
-        p!("[val/*:*/://\ntrue]" => [func!("val": (Bool(true)))]);
+        p!("[val\n:\ntrue]"      => [par![func!("val": (Bool(true)))]]);
+        p!("[val/*:*/://\ntrue]" => [par![func!("val": (Bool(true)))]]);
     }
 
     #[test]
@@ -963,24 +980,25 @@ mod tests {
         pval!("12e-3cm/1pt" => (Div(Len(Length::cm(12e-3)), Len(Length::pt(1.0)))));
 
         // Span of expression.
-        p!("[val: 1 + 3]" => [(0:0, 0:12, func!((0:1, 0:4, "val"): (
-            (0:6, 0:11, Expr::Add(
+        p!("[val: 1 + 3]" => [(0:0, 0:12, par![(0:0, 0:12, func!(
+            (0:1, 0:4, "val"): ((0:6, 0:11, Expr::Add(
                 Box::new(span_item!((0:6, 0:7, Num(1.0)))),
                 Box::new(span_item!((0:10, 0:11, Num(3.0)))),
-            ))
-        )))]);
+            )))
+        ))])]);
 
         // Unclosed string.
-        p!("[val: \"hello]" => [func!("val": (Str("hello]")), {})], [
+        p!("[val: \"hello]" => [par![func!("val": (Str("hello]")), {})]], [
             (0:13, 0:13, "expected quote"),
             (0:13, 0:13, "expected closing bracket"),
         ]);
 
         // Invalid, healed colors.
-        p!("[val: #12345]"     => [func!("val": (ColorHealed()))], [(0:6, 0:12, "invalid color")]);
-        p!("[val: #a5]"        => [func!("val": (ColorHealed()))], [(0:6, 0:9,  "invalid color")]);
-        p!("[val: #14b2ah]"    => [func!("val": (ColorHealed()))], [(0:6, 0:13, "invalid color")]);
-        p!("[val: #f075ff011]" => [func!("val": (ColorHealed()))], [(0:6, 0:16, "invalid color")]);
+        let healed = Expr::Color(RgbaColor::new_healed(0, 0, 0, 255));
+        p!("[val: #12345]"     => [par![func!("val": (healed.clone()))]], [(0:6, 0:12, "invalid color")]);
+        p!("[val: #a5]"        => [par![func!("val": (healed.clone()))]], [(0:6, 0:9,  "invalid color")]);
+        p!("[val: #14b2ah]"    => [par![func!("val": (healed.clone()))]], [(0:6, 0:13, "invalid color")]);
+        p!("[val: #f075ff011]" => [par![func!("val": (healed.clone()))]], [(0:6, 0:16, "invalid color")]);
     }
 
     #[test]
@@ -999,15 +1017,17 @@ mod tests {
         pval!("3/4*5" => (Mul(Div(Num(3.0), Num(4.0)), Num(5.0))));
 
         // Span of parenthesized expression contains parens.
-        p!("[val: (1)]" => [(0:0, 0:10, func!((0:1, 0:4, "val"): ((0:6, 0:9, Num(1.0)))))]);
+        p!("[val: (1)]" => [(0:0, 0:10, par![
+            (0:0, 0:10, func!((0:1, 0:4, "val"): ((0:6, 0:9, Num(1.0)))))
+        ])]);
 
         // Invalid expressions.
-        p!("[val: 4pt--]" => [func!("val": (Len(Length::pt(4.0))))], [
+        p!("[val: 4pt--]" => [par![func!("val": (Len(Length::pt(4.0))))]], [
             (0:10, 0:11, "dangling minus"),
             (0:6, 0:10, "missing right summand")
         ]);
         p!("[val: 3mm+4pt*]" =>
-            [func!("val": (Add(Len(Length::mm(3.0)), Len(Length::pt(4.0)))))],
+            [par![func!("val": (Add(Len(Length::mm(3.0)), Len(Length::pt(4.0)))))]],
             [(0:10, 0:14, "missing right factor")],
         );
     }
@@ -1026,19 +1046,19 @@ mod tests {
 
         // Invalid value.
         p!("[val: sound(\x07)]" =>
-            [func!("val": (named_tuple!("sound")), {})],
+            [par![func!("val": (named_tuple!("sound")), {})]],
             [(0:12, 0:13, "expected value, found invalid token")],
         );
 
         // Invalid tuple name.
         p!("[val: 👠(\"abc\", 13e-5)]" =>
-            [func!("val": (tuple!(Str("abc"), Num(13.0e-5))), {})],
+            [par![func!("val": (tuple!(Str("abc"), Num(13.0e-5))), {})]],
             [(0:6, 0:7, "expected argument, found invalid token")],
         );
 
         // Unclosed tuple.
         p!("[val: lang(中文]" =>
-            [func!("val": (named_tuple!("lang", Id("中文"))), {})],
+            [par![func!("val": (named_tuple!("lang", Id("中文"))), {})]],
             [(0:13, 0:13, "expected closing paren")],
         );
 
@@ -1059,18 +1079,18 @@ mod tests {
 
         // Invalid commas.
         p!("[val: (,)]" =>
-            [func!("val": (tuple!()), {})],
+            [par![func!("val": (tuple!()), {})]],
             [(0:7, 0:8, "expected value, found comma")],
         );
         p!("[val: (true false)]" =>
-            [func!("val": (tuple!(Bool(true), Bool(false))), {})],
+            [par![func!("val": (tuple!(Bool(true), Bool(false))), {})]],
             [(0:11, 0:11, "expected comma")],
         );
     }
 
     #[test]
     fn parse_objects() {
-        let val = || func!("val": (object! {}), {});
+        let val = || par![func!("val": (object! {}), {})];
 
         // Okay objects.
         pval!("{}" => (object! {}));
@@ -1078,11 +1098,11 @@ mod tests {
 
         // Unclosed object.
         p!("[val: {hello: world]" =>
-            [func!("val": (object! { "hello" => Id("world") }), {})],
+            [par![func!("val": (object! { "hello" => Id("world") }), {})]],
             [(0:19, 0:19, "expected closing brace")],
         );
         p!("[val: { a]" =>
-            [func!("val": (object! {}), {})],
+            [par![func!("val": (object! {}), {})]],
             [(0:9, 0:9, "expected colon"), (0:9, 0:9, "expected closing brace")],
         );
 
@@ -1098,25 +1118,25 @@ mod tests {
             (0:12, 0:17, "expected key, found bool"),
         ]);
         p!("[val: { a b:c }]" =>
-            [func!("val": (object! { "b" => Id("c") }), {})],
+            [par![func!("val": (object! { "b" => Id("c") }), {})]],
             [(0:9, 0:9, "expected colon")],
         );
 
         // Missing value.
         p!("[val: { key: : }]" => [val()], [(0:13, 0:14, "expected value, found colon")]);
         p!("[val: { key: , k: \"s\" }]" =>
-            [func!("val": (object! { "k" => Str("s") }), {})],
+            [par![func!("val": (object! { "k" => Str("s") }), {})]],
             [(0:13, 0:14, "expected value, found comma")],
         );
 
         // Missing comma, invalid token.
         p!("[val: left={ a: 2, b: false 🌎 }]" =>
-            [func!("val": (), {
+            [par![func!("val": (), {
                 "left" => object! {
                     "a" => Num(2.0),
                     "b" => Bool(false),
                 }
-            })],
+            })]],
             [(0:27, 0:27, "expected comma"),
              (0:28, 0:29, "expected key, found invalid token")],
         );
@@ -1140,19 +1160,19 @@ mod tests {
     fn parse_one_keyword_argument() {
         // Correct
         p!("[val: x=true]" =>
-            [func!("val": (), { "x" => Bool(true) })], [],
+            [par![func!("val": (), { "x" => Bool(true) })]], [],
             [(0:6, 0:7, ArgumentKey), (0:1, 0:4, ResolvedFunc)],
         );
 
         // Spacing around keyword arguments
         p!("\n [val: \n hi \n = /* //\n */ \"s\n\"]" =>
-            [S, func!("val": (), { "hi" => Str("s\n") })], [],
+            [par![S, func!("val": (), { "hi" => Str("s\n") })]], [],
             [(2:1, 2:3, ArgumentKey), (1:2, 1:5, ResolvedFunc)],
         );
 
         // Missing value
         p!("[val: x=]" =>
-            [func!("val")],
+            [par![func!("val")]],
             [(0:8, 0:8, "expected value")],
             [(0:6, 0:7, ArgumentKey), (0:1, 0:4, ResolvedFunc)],
         );
@@ -1161,7 +1181,8 @@ mod tests {
     #[test]
     fn parse_multiple_mixed_arguments() {
         p!("[val: 12pt, key=value]" =>
-            [func!("val": (Len(Length::pt(12.0))), { "key" => Id("value") })], [],
+            [par![func!("val": (Len(Length::pt(12.0))), { "key" => Id("value") })]],
+            [],
             [(0:12, 0:15, ArgumentKey), (0:1, 0:4, ResolvedFunc)],
         );
         pval!("a , x=\"b\" , c" => (Id("a"), Id("c")), { "x" => Str("b") });
@@ -1169,15 +1190,15 @@ mod tests {
 
     #[test]
     fn parse_invalid_values() {
-        p!("[val: )]"     => [func!("val")], [(0:6, 0:7, "expected argument, found closing paren")]);
-        p!("[val: }]"     => [func!("val")], [(0:6, 0:7, "expected argument, found closing brace")]);
-        p!("[val: :]"     => [func!("val")], [(0:6, 0:7, "expected argument, found colon")]);
-        p!("[val: ,]"     => [func!("val")], [(0:6, 0:7, "expected argument, found comma")]);
-        p!("[val: =]"     => [func!("val")], [(0:6, 0:7, "expected argument, found equals sign")]);
-        p!("[val: 🌎]"    => [func!("val")], [(0:6, 0:7, "expected argument, found invalid token")]);
-        p!("[val: 12ept]" => [func!("val")], [(0:6, 0:11, "expected argument, found invalid token")]);
+        p!("[val: )]"     => [par![func!("val")]], [(0:6, 0:7, "expected argument, found closing paren")]);
+        p!("[val: }]"     => [par![func!("val")]], [(0:6, 0:7, "expected argument, found closing brace")]);
+        p!("[val: :]"     => [par![func!("val")]], [(0:6, 0:7, "expected argument, found colon")]);
+        p!("[val: ,]"     => [par![func!("val")]], [(0:6, 0:7, "expected argument, found comma")]);
+        p!("[val: =]"     => [par![func!("val")]], [(0:6, 0:7, "expected argument, found equals sign")]);
+        p!("[val: 🌎]"    => [par![func!("val")]], [(0:6, 0:7, "expected argument, found invalid token")]);
+        p!("[val: 12ept]" => [par![func!("val")]], [(0:6, 0:11, "expected argument, found invalid token")]);
         p!("[val: [hi]]"  =>
-            [func!("val")],
+            [par![func!("val")]],
             [(0:6, 0:10, "expected argument, found function")],
             [(0:1, 0:4, ResolvedFunc)],
         );
@@ -1187,7 +1208,7 @@ mod tests {
     fn parse_invalid_key_value_pairs() {
         // Invalid keys.
         p!("[val: true=you]" =>
-            [func!("val": (Bool(true), Id("you")), {})],
+            [par![func!("val": (Bool(true), Id("you")), {})]],
             [(0:10, 0:10, "expected comma"),
              (0:10, 0:11, "expected argument, found equals sign")],
             [(0:1, 0:4, ResolvedFunc)],
@@ -1195,21 +1216,22 @@ mod tests {
 
         // Unexpected equals.
         p!("[box: z=y=4]" =>
-            [func!("box": (Num(4.0)), { "z" => Id("y") })],
+            [par![func!("box": (Num(4.0)), { "z" => Id("y") })]],
             [(0:9, 0:9, "expected comma"),
              (0:9, 0:10, "expected argument, found equals sign")],
         );
 
         // Invalid colon after keyable positional argument.
         p!("[val: key:12]" =>
-            [func!("val": (Id("key"), Num(12.0)), {})],
+            [par![func!("val": (Id("key"), Num(12.0)), {})]],
             [(0:9, 0:9, "expected comma"),
              (0:9, 0:10, "expected argument, found colon")],
             [(0:1, 0:4, ResolvedFunc)],
         );
 
         // Invalid colon after unkeyable positional argument.
-        p!("[val: true:12]" => [func!("val": (Bool(true), Num(12.0)), {})],
+        p!("[val: true:12]" =>
+            [par![func!("val": (Bool(true), Num(12.0)), {})]],
             [(0:10, 0:10, "expected comma"),
              (0:10, 0:11, "expected argument, found colon")],
             [(0:1, 0:4, ResolvedFunc)],
@@ -1220,33 +1242,33 @@ mod tests {
     fn parse_invalid_commas() {
         // Missing commas.
         p!("[val: 1pt 1]" =>
-            [func!("val": (Len(Length::pt(1.0)), Num(1.0)), {})],
+            [par![func!("val": (Len(Length::pt(1.0)), Num(1.0)), {})]],
             [(0:9, 0:9, "expected comma")],
         );
         p!(r#"[val: _"s"]"# =>
-            [func!("val": (Id("_"), Str("s")), {})],
+            [par![func!("val": (Id("_"), Str("s")), {})]],
             [(0:7, 0:7, "expected comma")],
         );
 
         // Unexpected commas.
-        p!("[val:,]" => [func!("val")], [(0:5, 0:6, "expected argument, found comma")]);
-        p!("[val: key=,]" => [func!("val")], [(0:10, 0:11, "expected value, found comma")]);
+        p!("[val:,]" => [par![func!("val")]], [(0:5, 0:6, "expected argument, found comma")]);
+        p!("[val: key=,]" => [par![func!("val")]], [(0:10, 0:11, "expected value, found comma")]);
         p!("[val:, true]" =>
-            [func!("val": (Bool(true)), {})],
+            [par![func!("val": (Bool(true)), {})]],
             [(0:5, 0:6, "expected argument, found comma")],
         );
     }
 
     #[test]
     fn parse_bodies() {
-        p!("[val][Hi]" => [func!("val"; [T("Hi")])]);
+        p!("[val][Hi]" => [par![func!("val"; [par![T("Hi")]])]]);
         p!("[val:*][*Hi*]" =>
-            [func!("val"; [Bold, T("Hi"), Bold])],
+            [par![func!("val"; [par![Bold, T("Hi"), Bold]])]],
             [(0:5, 0:6, "expected argument, found star")],
         );
         // Errors in bodies.
         p!(" [val][ */ ]" =>
-            [S, func!("val"; [S, S])],
+            [par![S, func!("val"; [par![S, S]])]],
             [(0:8, 0:10, "unexpected end of block comment")],
         );
     }
@@ -1255,40 +1277,47 @@ mod tests {
     fn parse_spanned_functions() {
         // Space before function
         p!(" [val]" =>
-            [(0:0, 0:1, S), (0:1, 0:6, func!((0:1, 0:4, "val")))], [],
+            [(0:0, 0:6, par![(0:0, 0:1, S), (0:1, 0:6, func!((0:1, 0:4, "val")))])],
+            [],
             [(0:2, 0:5, ResolvedFunc)],
         );
 
         // Newline before function
-        p!(" \n\r\n[val]" =>
-            [(0:0, 2:0, Parbreak), (2:0, 2:5, func!((0:1, 0:4, "val")))], [],
+        p!("a \n\r\n[val]" =>
+            [
+                (0:0, 0:1, par![(0:0, 0:1, T("a"))]),
+                (2:0, 2:5, par![(2:0, 2:5, func!((0:1, 0:4, "val")))]),
+            ],
+            [],
             [(2:1, 2:4, ResolvedFunc)],
         );
 
         // Content before function
         p!("hello [val][world] 🌎" =>
-            [
+            [(0:0, 0:20, par![
                 (0:0, 0:5, T("hello")),
                 (0:5, 0:6, S),
-                (0:6, 0:18, func!((0:1, 0:4, "val"); [(0:6, 0:11, T("world"))])),
+                (0:6, 0:18, func!((0:1, 0:4, "val"); [
+                    (0:6, 0:11, par![(0:6, 0:11, T("world"))])
+                ])),
                 (0:18, 0:19, S),
                 (0:19, 0:20, T("🌎"))
-            ],
+            ])],
             [],
             [(0:7, 0:10, ResolvedFunc)],
         );
 
         // Nested function
         p!(" [val][\nbody[ box]\n ]" =>
-            [
+            [(0:0, 2:2, par![
                 (0:0, 0:1, S),
-                (0:1, 2:2, func!((0:1, 0:4, "val"); [
+                (0:1, 2:2, func!((0:1, 0:4, "val"); [(0:6, 2:1, par![
                     (0:6, 1:0, S),
                     (1:0, 1:4, T("body")),
                     (1:4, 1:10, func!((0:2, 0:5, "box"))),
                     (1:10, 2:1, S),
-                ]))
-            ],
+                ])]))
+            ])],
             [],
             [(0:2, 0:5, ResolvedFunc), (1:6, 1:9, ResolvedFunc)],
         );
diff --git a/src/syntax/test.rs b/src/syntax/test.rs
index db7a2de2e..9faa7f233 100644
--- a/src/syntax/test.rs
+++ b/src/syntax/test.rs
@@ -86,22 +86,23 @@ pub trait SpanlessEq<Rhs = Self> {
 }
 
 impl SpanlessEq for SyntaxNode {
-    fn spanless_eq(&self, other: &SyntaxNode) -> bool {
+    fn spanless_eq(&self, other: &Self) -> bool {
         fn downcast<'a>(func: &'a (dyn DynamicNode + 'static)) -> &'a DebugFn {
             func.downcast::<DebugFn>().expect("not a debug fn")
         }
 
         match (self, other) {
-            (SyntaxNode::Dyn(a), SyntaxNode::Dyn(b)) => {
+            (Self::Dyn(a), Self::Dyn(b)) => {
                 downcast(a.as_ref()).spanless_eq(downcast(b.as_ref()))
             }
+            (Self::Par(a), Self::Par(b)) => a.spanless_eq(b),
             (a, b) => a == b,
         }
     }
 }
 
 impl SpanlessEq for DebugFn {
-    fn spanless_eq(&self, other: &DebugFn) -> bool {
+    fn spanless_eq(&self, other: &Self) -> bool {
         self.header.spanless_eq(&other.header)
             && self.body.spanless_eq(&other.body)
     }
@@ -132,7 +133,7 @@ impl SpanlessEq for FuncArg {
 }
 
 impl SpanlessEq for Expr {
-    fn spanless_eq(&self, other: &Expr) -> bool {
+    fn spanless_eq(&self, other: &Self) -> bool {
         match (self, other) {
             (Expr::Tuple(a), Expr::Tuple(b)) => a.spanless_eq(b),
             (Expr::NamedTuple(a), Expr::NamedTuple(b)) => a.spanless_eq(b),
@@ -148,20 +149,20 @@ impl SpanlessEq for Expr {
 }
 
 impl SpanlessEq for Tuple {
-    fn spanless_eq(&self, other: &Tuple) -> bool {
+    fn spanless_eq(&self, other: &Self) -> bool {
         self.0.spanless_eq(&other.0)
     }
 }
 
 impl SpanlessEq for NamedTuple {
-    fn spanless_eq(&self, other: &NamedTuple) -> bool {
+    fn spanless_eq(&self, other: &Self) -> bool {
         self.name.v == other.name.v
             && self.tuple.v.spanless_eq(&other.tuple.v)
     }
 }
 
 impl SpanlessEq for Object {
-    fn spanless_eq(&self, other: &Object) -> bool {
+    fn spanless_eq(&self, other: &Self) -> bool {
         self.0.spanless_eq(&other.0)
     }
 }
@@ -173,20 +174,20 @@ impl SpanlessEq for Pair {
 }
 
 impl<T: SpanlessEq> SpanlessEq for Vec<T> {
-    fn spanless_eq(&self, other: &Vec<T>) -> bool {
+    fn spanless_eq(&self, other: &Self) -> bool {
         self.len() == other.len()
             && self.iter().zip(other).all(|(x, y)| x.spanless_eq(&y))
     }
 }
 
 impl<T: SpanlessEq> SpanlessEq for Spanned<T> {
-    fn spanless_eq(&self, other: &Spanned<T>) -> bool {
+    fn spanless_eq(&self, other: &Self) -> bool {
         self.v.spanless_eq(&other.v)
     }
 }
 
 impl<T: SpanlessEq> SpanlessEq for Box<T> {
-    fn spanless_eq(&self, other: &Box<T>) -> bool {
+    fn spanless_eq(&self, other: &Self) -> bool {
         (&**self).spanless_eq(&**other)
     }
 }
diff --git a/src/syntax/tree.rs b/src/syntax/tree.rs
index 28997e7cb..73734d28d 100644
--- a/src/syntax/tree.rs
+++ b/src/syntax/tree.rs
@@ -14,20 +14,20 @@ pub type SyntaxTree = SpanVec<SyntaxNode>;
 #[derive(Debug, Clone)]
 pub enum SyntaxNode {
     /// Whitespace containing less than two newlines.
-    Space,
-    /// Whitespace with more than two newlines.
-    Parbreak,
+    Spacing,
     /// A forced line break.
     Linebreak,
-    /// Plain text.
-    Text(String),
-    /// Lines of raw text.
-    Raw(Vec<String>),
     /// Italics were enabled / disabled.
     ToggleItalic,
     /// Bolder was enabled / disabled.
     ToggleBolder,
-    /// A dynamic node, create through function invocations in source code.
+    /// Plain text.
+    Text(String),
+    /// Lines of raw text.
+    Raw(Vec<String>),
+    /// A paragraph of child nodes.
+    Par(SyntaxTree),
+    /// A dynamic node, created through function invocations in source code.
     Dyn(Box<dyn DynamicNode>),
 }
 
@@ -35,13 +35,13 @@ impl PartialEq for SyntaxNode {
     fn eq(&self, other: &SyntaxNode) -> bool {
         use SyntaxNode::*;
         match (self, other) {
-            (Space, Space) => true,
-            (Parbreak, Parbreak) => true,
+            (Spacing, Spacing) => true,
             (Linebreak, Linebreak) => true,
-            (Text(a), Text(b)) => a == b,
-            (Raw(a), Raw(b)) => a == b,
             (ToggleItalic, ToggleItalic) => true,
             (ToggleBolder, ToggleBolder) => true,
+            (Text(a), Text(b)) => a == b,
+            (Raw(a), Raw(b)) => a == b,
+            (Par(a), Par(b)) => a == b,
             (Dyn(a), Dyn(b)) => a == b,
             _ => false,
         }