diff --git a/lib/coffeescript/lexer.js b/lib/coffeescript/lexer.js
index 884f9881..3a250189 100644
--- a/lib/coffeescript/lexer.js
+++ b/lib/coffeescript/lexer.js
@@ -672,9 +672,7 @@
       if (dent) {
         this.outdebt -= moveOut;
       }
-      while (this.value() === ';') {
-        this.tokens.pop();
-      }
+      this.suppressSemicolons();
       if (!(this.tag() === 'TERMINATOR' || noNewlines)) {
         this.token('TERMINATOR', '\n', outdentLength, 0);
       }
@@ -703,9 +701,7 @@
 
     // Generate a newline token. Consecutive newlines get merged together.
     newlineToken(offset) {
-      while (this.value() === ';') {
-        this.tokens.pop();
-      }
+      this.suppressSemicolons();
       if (this.tag() !== 'TERMINATOR') {
         this.token('TERMINATOR', '\n', offset, 0);
       }
@@ -844,7 +840,7 @@
     // here. `;` and newlines are both treated as a `TERMINATOR`, we distinguish
     // parentheses that indicate a method call from regular parentheses, and so on.
     literalToken() {
-      var match, message, origin, prev, ref, ref1, ref2, ref3, skipToken, tag, token, value;
+      var match, message, origin, prev, ref, ref1, ref2, ref3, ref4, skipToken, tag, token, value;
       if (match = OPERATOR.exec(this.chunk)) {
         [value] = match;
         if (CODE.test(value)) {
@@ -884,6 +880,9 @@
         this.exportSpecifierList = false;
       }
       if (value === ';') {
+        if (ref2 = prev != null ? prev[0] : void 0, indexOf.call(['=', ...UNFINISHED], ref2) >= 0) {
+          this.error('unexpected ;');
+        }
         this.seenFor = this.seenImport = this.seenExport = false;
         tag = 'TERMINATOR';
       } else if (value === '*' && (prev != null ? prev[0] : void 0) === 'EXPORT') {
@@ -903,12 +902,12 @@
       } else if (value === '?' && (prev != null ? prev.spaced : void 0)) {
         tag = 'BIN?';
       } else if (prev) {
-        if (value === '(' && !prev.spaced && (ref2 = prev[0], indexOf.call(CALLABLE, ref2) >= 0)) {
+        if (value === '(' && !prev.spaced && (ref3 = prev[0], indexOf.call(CALLABLE, ref3) >= 0)) {
           if (prev[0] === '?') {
             prev[0] = 'FUNC_EXIST';
           }
           tag = 'CALL_START';
-        } else if (value === '[' && (((ref3 = prev[0], indexOf.call(INDEXABLE, ref3) >= 0) && !prev.spaced) || (prev[0] === '::'))) { // `.prototype` can’t be a method you can call.
+        } else if (value === '[' && (((ref4 = prev[0], indexOf.call(INDEXABLE, ref4) >= 0) && !prev.spaced) || (prev[0] === '::'))) { // `.prototype` can’t be a method you can call.
           tag = 'INDEX_START';
           switch (prev[0]) {
             case '?':
@@ -1407,6 +1406,20 @@
       return `${options.delimiter}${body}${options.delimiter}`;
     }
 
+    suppressSemicolons() {
+      var ref, ref1, results;
+      results = [];
+      while (this.value() === ';') {
+        this.tokens.pop();
+        if (ref = (ref1 = this.prev()) != null ? ref1[0] : void 0, indexOf.call(['=', ...UNFINISHED], ref) >= 0) {
+          results.push(this.error('unexpected ;'));
+        } else {
+          results.push(void 0);
+        }
+      }
+      return results;
+    }
+
     // Throws an error at either a given offset from the current chunk or at the
     // location of a token (`token[2]`).
     error(message, options = {}) {
diff --git a/src/lexer.coffee b/src/lexer.coffee
index b31ac13d..f3b2509a 100644
--- a/src/lexer.coffee
+++ b/src/lexer.coffee
@@ -509,7 +509,7 @@ exports.Lexer = class Lexer
         @token 'OUTDENT', moveOut, 0, outdentLength
         moveOut -= dent
     @outdebt -= moveOut if dent
-    @tokens.pop() while @value() is ';'
+    @suppressSemicolons()
 
     @token 'TERMINATOR', '\n', outdentLength, 0 unless @tag() is 'TERMINATOR' or noNewlines
     @indent = decreasedIndent
@@ -527,7 +527,7 @@ exports.Lexer = class Lexer
 
   # Generate a newline token. Consecutive newlines get merged together.
   newlineToken: (offset) ->
-    @tokens.pop() while @value() is ';'
+    @suppressSemicolons()
     @token 'TERMINATOR', '\n', offset, 0 unless @tag() is 'TERMINATOR'
     this
 
@@ -662,6 +662,7 @@ exports.Lexer = class Lexer
       @exportSpecifierList = no
 
     if value is ';'
+      @error 'unexpected ;' if prev?[0] in ['=', UNFINISHED...]
       @seenFor = @seenImport = @seenExport = no
       tag = 'TERMINATOR'
     else if value is '*' and prev?[0] is 'EXPORT'
@@ -1053,6 +1054,11 @@ exports.Lexer = class Lexer
       when other     then (if options.double then "\\#{other}" else other)
     "#{options.delimiter}#{body}#{options.delimiter}"
 
+  suppressSemicolons: ->
+    while @value() is ';'
+      @tokens.pop()
+      @error 'unexpected ;' if @prev()?[0] in ['=', UNFINISHED...]
+
   # Throws an error at either a given offset from the current chunk or at the
   # location of a token (`token[2]`).
   error: (message, options = {}) ->
diff --git a/test/error_messages.coffee b/test/error_messages.coffee
index d5aeda60..640e46c4 100644
--- a/test/error_messages.coffee
+++ b/test/error_messages.coffee
@@ -1759,3 +1759,12 @@ test "#3199: error message for throw indented comprehension", ->
       x for x in [1, 2, 3]
       ^
   '''
+
+test "#3098: suppressed newline should be unsuppressed by semicolon", ->
+  assertErrorFormat '''
+    a = ; 5
+  ''', '''
+    [stdin]:1:5: error: unexpected ;
+    a = ; 5
+        ^
+  '''