diff --git a/main.go b/main.go
index 19d0ccb..44844b9 100644
--- a/main.go
+++ b/main.go
@@ -262,6 +262,31 @@ func (m *model) handleDigKey(msg tea.KeyMsg) (tea.Model, tea.Cmd) {
 			}
 		}
 
+	case key.Matches(msg, textinput.DefaultKeyMap.WordBackward):
+		value := m.digInput.Value()
+		pth, ok := jsonpath.Split(value[0:m.digInput.Position()])
+		if ok {
+			if len(pth) > 0 {
+				pth = pth[:len(pth)-1]
+				m.digInput.SetCursor(len(jsonpath.Join(pth)))
+			} else {
+				m.digInput.CursorStart()
+			}
+		}
+
+	case key.Matches(msg, textinput.DefaultKeyMap.WordForward):
+		value := m.digInput.Value()
+		fullPath, ok1 := jsonpath.Split(value)
+		pth, ok2 := jsonpath.Split(value[0:m.digInput.Position()])
+		if ok1 && ok2 {
+			if len(pth) < len(fullPath) {
+				pth = append(pth, fullPath[len(pth)])
+				m.digInput.SetCursor(len(jsonpath.Join(pth)))
+			} else {
+				m.digInput.CursorEnd()
+			}
+		}
+
 	default:
 		if key.Matches(msg, key.NewBinding(key.WithKeys("."))) {
 			m.digInput.SetValue(m.cursorPath())
@@ -752,7 +777,7 @@ func (m *model) cursorPath() string {
 			if at.key != nil {
 				quoted := string(at.key)
 				unquoted, err := strconv.Unquote(quoted)
-				if err == nil && identifier.MatchString(unquoted) {
+				if err == nil && jsonpath.Identifier.MatchString(unquoted) {
 					path = "." + unquoted + path
 				} else {
 					path = "[" + quoted + "]" + path
diff --git a/node.go b/node.go
index eb90b62..2925b0d 100644
--- a/node.go
+++ b/node.go
@@ -2,6 +2,8 @@ package main
 
 import (
 	"strconv"
+
+	jsonpath "github.com/antonmedv/fx/path"
 )
 
 type node struct {
@@ -177,7 +179,7 @@ func (n *node) paths(prefix string, paths *[]string, nodes *[]*node) {
 		if it.key != nil {
 			quoted := string(it.key)
 			unquoted, err := strconv.Unquote(quoted)
-			if err == nil && identifier.MatchString(unquoted) {
+			if err == nil && jsonpath.Identifier.MatchString(unquoted) {
 				path = prefix + "." + unquoted
 			} else {
 				path = prefix + "[" + quoted + "]"
diff --git a/path/path.go b/path/path.go
index 602f448..3dfe21e 100644
--- a/path/path.go
+++ b/path/path.go
@@ -1,6 +1,7 @@
 package path
 
 import (
+	"regexp"
 	"strconv"
 	"unicode"
 )
@@ -174,3 +175,22 @@ func Split(p string) ([]any, bool) {
 func isProp(ch rune) bool {
 	return unicode.IsLetter(ch) || unicode.IsDigit(ch) || ch == '_' || ch == '$'
 }
+
+var Identifier = regexp.MustCompile(`^[a-zA-Z_][a-zA-Z0-9_]*$`)
+
+func Join(path []any) string {
+	s := ""
+	for _, v := range path {
+		switch v := v.(type) {
+		case string:
+			if Identifier.MatchString(v) {
+				s += "." + v
+			} else {
+				s += "[" + strconv.Quote(v) + "]"
+			}
+		case int:
+			s += "[" + strconv.Itoa(v) + "]"
+		}
+	}
+	return s
+}
diff --git a/path/path_test.go b/path/path_test.go
index 09c4177..9360ed2 100644
--- a/path/path_test.go
+++ b/path/path_test.go
@@ -135,3 +135,52 @@ func Test_SplitPath_negative(t *testing.T) {
 		})
 	}
 }
+
+func TestJoin(t *testing.T) {
+	tests := []struct {
+		input []any
+		want  string
+	}{
+		{
+			input: []any{},
+			want:  "",
+		},
+		{
+			input: []any{"foo"},
+			want:  ".foo",
+		},
+		{
+			input: []any{"foo", "bar"},
+			want:  ".foo.bar",
+		},
+		{
+			input: []any{"foo", 42},
+			want:  ".foo[42]",
+		},
+		{
+			input: []any{"foo", "bar", 42},
+			want:  ".foo.bar[42]",
+		},
+		{
+			input: []any{"foo", "bar", 42, "baz"},
+			want:  ".foo.bar[42].baz",
+		},
+		{
+			input: []any{"foo", "bar", 42, "baz", 1},
+			want:  ".foo.bar[42].baz[1]",
+		},
+		{
+			input: []any{"foo", "bar", 42, "baz", 1, "qux"},
+			want:  ".foo.bar[42].baz[1].qux",
+		},
+		{
+			input: []any{"foo bar"},
+			want:  "[\"foo bar\"]",
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.want, func(t *testing.T) {
+			require.Equal(t, tt.want, path.Join(tt.input))
+		})
+	}
+}
diff --git a/utils.go b/utils.go
index 5831410..af354a1 100644
--- a/utils.go
+++ b/utils.go
@@ -1,12 +1,9 @@
 package main
 
 import (
-	"regexp"
 	"strings"
 )
 
-var identifier = regexp.MustCompile(`^[a-zA-Z_][a-zA-Z0-9_]*$`)
-
 func isHexDigit(ch byte) bool {
 	return (ch >= '0' && ch <= '9') || (ch >= 'a' && ch <= 'f') || (ch >= 'A' && ch <= 'F')
 }