diff --git a/lib/regexp_parser/scanner/scanner.rl b/lib/regexp_parser/scanner/scanner.rl
index d14146fe..82ca9073 100644
--- a/lib/regexp_parser/scanner/scanner.rl
+++ b/lib/regexp_parser/scanner/scanner.rl
@@ -128,7 +128,7 @@
   utf8_4_byte           = (0xf0..0xf4 0x80..0xbf 0x80..0xbf 0x80..0xbf);
 
   non_literal_escape    = char_type_char | anchor_char | escaped_ascii |
-                          group_ref | keep_mark | [xucCM];
+                          keep_mark | [xucCM];
 
   non_set_escape        = (anchor_char - 'b') | group_ref | keep_mark |
                           multi_codepoint_char_type | [0-9cCM];
diff --git a/spec/scanner/escapes_spec.rb b/spec/scanner/escapes_spec.rb
index 8579c2ca..9f84f4a0 100644
--- a/spec/scanner/escapes_spec.rb
+++ b/spec/scanner/escapes_spec.rb
@@ -13,6 +13,10 @@
 
   include_examples 'scan', 'c\qt',            1 => [:escape,  :literal,          '\q',             1,  3]
 
+  # these incomplete ref/call sequences are treated as literal escapes by Ruby
+  include_examples 'scan', 'c\gt',            1 => [:escape,  :literal,          '\g',             1,  3]
+  include_examples 'scan', 'c\kt',            1 => [:escape,  :literal,          '\k',             1,  3]
+
   include_examples 'scan', 'a\012c',          1 => [:escape,  :octal,            '\012',           1,  5]
   include_examples 'scan', 'a\0124',          1 => [:escape,  :octal,            '\012',           1,  5]
   include_examples 'scan', '\712+7',          0 => [:escape,  :octal,            '\712',           0,  4]