diff --git a/mathics/builtin/strings.py b/mathics/builtin/strings.py
index eddfb740f8..49dafa8612 100644
--- a/mathics/builtin/strings.py
+++ b/mathics/builtin/strings.py
@@ -1652,7 +1652,7 @@ class ToString(Builtin):
     """
 
     options = {
-        "CharacterEncoding": '"Unicode"',
+        "CharacterEncoding": "$SystemCharacterEncoding",
         "FormatType": "OutputForm",
         "NumberMarks": "$NumberMarks",
         "PageHeight": "Infinity",
@@ -1668,6 +1668,9 @@ def apply_default(self, value, evaluation, options):
     def apply_form(self, value, form, evaluation, options):
         "ToString[value_, form_, OptionsPattern[ToString]]"
         encoding = options["System`CharacterEncoding"]
+        if not isinstance(encoding, String) or encoding.value not in _encodings:
+            evaluation.message("General", "charcode", encoding)
+            encoding = Symbol("$SystemCharacterEncoding")
         text = value.format(evaluation, form.get_name(), encoding=encoding)
         text = text.boxes_to_text(evaluation=evaluation)
         return String(text)
diff --git a/mathics/core/evaluation.py b/mathics/core/evaluation.py
index ac988ed659..71b08778dc 100644
--- a/mathics/core/evaluation.py
+++ b/mathics/core/evaluation.py
@@ -13,7 +13,8 @@
 from mathics_scanner import TranslateError
 
 from mathics import settings
-from mathics.core.expression import ensure_context, KeyComparable, SymbolAborted, SymbolList, SymbolNull
+from mathics.core.expression import ensure_context, KeyComparable, SymbolAborted, SymbolList, SymbolNull, wl_strings_to_specific_encoding
+from mathics.settings import SYSTEM_CHARACTER_ENCODING
 
 FORMATS = [
     "StandardForm",
@@ -228,7 +229,12 @@ def display(self, data, metadata):
 
 class Evaluation(object):
     def __init__(
-        self, definitions=None, output=None, format="text", catch_interrupt=True
+        self,
+        definitions=None,
+        output=None,
+        format="text",
+        catch_interrupt=True,
+        default_encoding=SYSTEM_CHARACTER_ENCODING,
     ) -> None:
         from mathics.core.definitions import Definitions
         from mathics.core.expression import Symbol
@@ -249,6 +255,7 @@ def __init__(
         self.quiet_all = False
         self.format = format
         self.catch_interrupt = catch_interrupt
+        self.encoding = default_encoding
 
         self.SymbolNull = SymbolNull
 
@@ -510,6 +517,7 @@ def message(self, symbol, tag, *args) -> None:
             Expression("StringForm", text, *(from_python(arg) for arg in args)), "text"
         )
 
+        text = wl_strings_to_specific_encoding(text, encoding=self.encoding)
         self.out.append(Message(symbol_shortname, tag, text))
         self.output.out(self.out[-1])
 
diff --git a/mathics/core/expression.py b/mathics/core/expression.py
index 9a55368d0e..23d6191c37 100644
--- a/mathics/core/expression.py
+++ b/mathics/core/expression.py
@@ -17,6 +17,20 @@
 from mathics.core.convert import sympy_symbol_prefix, SympyExpression
 import base64
 
+# In the future, this function should be replaced by one
+# handling different encodings.
+from mathics_scanner.characters import replace_wl_with_plain_text
+
+
+def wl_strings_to_specific_encoding(string, enc=None):
+    """
+    This function takes a string defined in the default
+    unicode WL encoding, and translate it to the corresponding
+    codepage. By now it is just an envelopment for
+    replace_wl_with_plain_text
+    """
+    return replace_wl_with_plain_text(string, enc in ("UTF8", "UTF-8"))
+
 # Imperical number that seems to work.
 # We have to be able to match mpmath values with sympy values
 COMPARE_PREC = 50
@@ -1969,7 +1983,8 @@ def do_copy(self) -> "Symbol":
         return Symbol(self.name)
 
     def boxes_to_text(self, **options) -> str:
-        return str(self.name)
+        encoding = options.get("encoding", "UTF-8")
+        return wl_strings_to_specific_encoding(str(self.name), encoding)
 
     def atom_to_boxes(self, f, evaluation) -> "String":
         return String(evaluation.definitions.shorten_name(self.name))
@@ -2825,7 +2840,8 @@ def boxes_to_text(self, show_string_characters=False, **options) -> str:
         ):
             value = value[1:-1]
 
-        return value
+        encoding = options.get("encoding", "UTF-8")
+        return wl_strings_to_specific_encoding(value, encoding)
 
     def boxes_to_mathml(self, show_string_characters=False, **options) -> str:
         from mathics.core.parser import is_symbol_name