diff --git a/framework/pyactr/pyactr.go b/framework/pyactr/pyactr.go
index a98fb05b..924db0e2 100644
--- a/framework/pyactr/pyactr.go
+++ b/framework/pyactr/pyactr.go
@@ -116,6 +116,12 @@ func (p *PyACTR) WriteModel(path, initialGoal string) (outputFileName string, er
 
 	p.outputAuthors()
 
+	if p.model.HasPrintStatement() {
+		// We use csv to parse the print text we are generating.
+		// This is just simpler than writing it ourselves (i.e. handling "foo, bar ", 66).
+		p.Writeln("import csv")
+	}
+
 	p.Writeln("import pyactr as actr")
 
 	if p.model.HasPrintStatement() {
@@ -137,20 +143,6 @@ func (p *PyACTR) WriteModel(path, initialGoal string) (outputFileName string, er
 
 	p.Write("%s = actr.ACTRModel(%s)\n\n", p.className, strings.Join(additionalInit, ", "))
 
-	if p.model.HasPrintStatement() {
-		p.Writeln(`
-# Monkey patch Buffer to add a new method.
-# Currently we can only output simple strings.
-def print_text(*args):
-    text = ''.join(args[1:])
-    text = text.strip("'")
-    print(text)
-
-
-Buffer.print_text = print_text
-`)
-	}
-
 	// chunks
 	for _, chunk := range p.model.Chunks {
 		if chunk.IsInternal() {
@@ -218,13 +210,6 @@ Buffer.print_text = print_text
 		if production.DoStatements != nil {
 			for _, statement := range production.DoStatements {
 				if statement.Print != nil {
-					if !onlyStrings(statement.Print.Values) {
-						warning := fmt.Sprintf("Warning: ('%s') pyactr currently cannot print variables from productions", production.Name)
-						warnings = append(warnings, warning)
-
-						continue
-					}
-
 					numPrintStatements++
 					if numPrintStatements > 1 {
 						warning := fmt.Sprintf("Warning: ('%s') pyactr currently only supports one print statement per production", production.Name)
@@ -233,7 +218,7 @@ Buffer.print_text = print_text
 					}
 				}
 
-				p.outputStatement(statement)
+				p.outputStatement(production, statement)
 			}
 		}
 
@@ -250,12 +235,80 @@ Buffer.print_text = print_text
 
 	p.Writeln("")
 
+	if p.model.HasPrintStatement() {
+		// We add some python code to monkey patch the model and the buffer.
+		// This lets us output strings, numbers, and the contents of buffer slots.
+
+		p.Writeln(`
+# Monkey patch ACTRModel to add a new method.
+def get_buffer(self, buffer_name: str) -> Buffer:
+	if buffer_name == 'goal':
+		return self._ACTRModel__buffers['goal']
+	if buffer_name == 'retrieval':
+		return self._ACTRModel__buffers["retrieval"]
+
+	print('ERROR: Buffer \'' + buffer_name + '\' not found')
+
+
+actr.ACTRModel.get_buffer = get_buffer
+
+
+# Monkey patch Buffer to add a new methods.
+def get_slot_contents(self, buffer_name: str, slot_name: str) -> str:
+	if self._data:
+		chunk = self._data.copy().pop()
+	else:
+		chunk = None
+	try:
+		return str(getattr(chunk, slot_name))
+	except AttributeError:
+		print('ERROR: no slot named \'' + slot_name +
+			  '\' in buffer \'' + buffer_name + '\'')
+
+	return ''
+
+
+def print_text(*args):
+	text = ''.join(args[1:]).strip('"')
+	output = ''  # build up our output in this buffer
+
+	for itemlist in csv.reader([text]):
+		for item in itemlist:
+			item = item.strip(' ')
+
+			# Handle string
+			if item[0] == '\'' or item[0] == '"':
+				output += item[1:-1]
+			else:
+				# Handle number
+				try:
+					float(item)
+					output += item
+				except ValueError:
+					# If we are here, we should have a buffer.slotname
+					ids = item.split('.')
+					if len(ids) != 2:
+						print(
+							'ERROR: expected <buffer>.<slot_name>, found \'' + item + '\'')
+					else:
+						buffer = %s.get_buffer(ids[0])
+						output += buffer.get_slot_contents(ids[0], ids[1])
+
+	print(output)
+
+
+Buffer.get_slot_contents = get_slot_contents
+Buffer.print_text = print_text
+`, p.className)
+	}
+
 	// ...add our code to run
+	p.Writeln("# Main")
 	p.Writeln("if __name__ == '__main__':")
 	p.Writeln("\tsim = %s.simulation()", p.className)
 	p.Writeln("\tsim.run()")
 	// TODO: Add some intelligent output when logging level is info or detail
-	p.Writeln("\tif goal.test_buffer('full') == True:")
+	p.Writeln("\tif goal.test_buffer('full') is True:")
 	p.Writeln("\t\tprint('final goal: ' + str(goal.pop()))")
 
 	return
@@ -335,7 +388,7 @@ func addPatternSlot(tabbedItems *framework.KeyValueList, slotName string, patter
 	}
 }
 
-func (p *PyACTR) outputStatement(s *actr.Statement) {
+func (p *PyACTR) outputStatement(production *actr.Production, s *actr.Statement) {
 	if s.Set != nil {
 		buffer := s.Set.Buffer
 		bufferName := buffer.GetName()
@@ -367,9 +420,24 @@ func (p *PyACTR) outputStatement(s *actr.Statement) {
 		p.Writeln("\t+retrieval>")
 		p.outputPattern(s.Recall.Pattern, 2)
 	} else if s.Print != nil {
+		// Using "goal" here is arbitrary because of the way we monkey patch the python code.
+		// Our "print_text" statement handles its own formatting and lookup.
 		p.Writeln("\t!goal>")
-		values := framework.PythonValuesToStrings(s.Print.Values, true)
-		p.Writeln("\t\tprint_text %s", strings.Join(values, " "))
+
+		str := make([]string, len(*s.Print.Values))
+
+		for index, val := range *s.Print.Values {
+			if val.Var != nil {
+				varIndex := production.VarIndexMap[*val.Var]
+				str[index] = fmt.Sprintf("%s.%s", varIndex.Buffer.GetName(), varIndex.SlotName)
+			} else if val.Str != nil {
+				str[index] = fmt.Sprintf("'%s'", *val.Str)
+			} else if val.Number != nil {
+				str[index] = *val.Number
+			}
+		}
+
+		p.Writeln("\t\tprint_text \"%s\"", strings.Join(str, ", "))
 	} else if s.Clear != nil {
 		for _, name := range s.Clear.BufferNames {
 			p.Writeln("\t~%s>", name)
@@ -377,17 +445,6 @@ func (p *PyACTR) outputStatement(s *actr.Statement) {
 	}
 }
 
-func onlyStrings(values *[]*actr.Value) bool {
-	for _, v := range *values {
-		if v.Var != nil {
-			return false
-		}
-		// v.ID should not be possible because of validation
-	}
-
-	return true
-}
-
 // removeWarning will remove the long warning whenever pyactr is run without tkinter.
 func removeWarning(text []byte) []byte {
 	str := string(text)