Literature Review of articles inSpeechVoiceReadingDeafnessLanguage

What is that little voice inside my head? Inner speech phenomenology, its role in cognitive performance, and its relation to self-monitoring

Article · Literature Review · January 2014with1,529 Reads
DOI: 10.1016/j.bbr.2013.12.034 · Source: PubMed
Abstract
The little voice inside our head, or inner speech, is a common everyday experience. It plays a central role in human consciousness at the interplay of language and thought. An impressive host of research works has been carried out on inner speech these last fifty years. Here we first describe the phenomenology of inner speech by examining five issues: common behavioural and cerebral correlates with overt speech, different types of inner speech (wilful verbal thought generation and verbal mind wandering), presence of inner speech in reading and in writing, inner signing and voice-hallucinations in deaf people. Secondly, we review the role of inner speech in cognitive performance (i.e. enhancement vs. perturbation). Finally, we consider agency in inner speech and how our inner voice is known to be self-generated and not produced by someone else.
1 Figures
Behavioural
Brain
Research
261 (2014) 220–
239
Contents
lists
available
at
ScienceDirect
Behavioural
Brain
Research
j
ourna
l
h
o
mepa
ge:
www.elsevier.com/locate/bbr
Review
What
is
that
little
voice
inside
my
head?
Inner
speech
phenomenology,
its
role
in
cognitive
performance,
and
its
relation
to
self-monitoring
M.
Perrone-Bertolotti
a
,b,c,d,e,
,
L.
Rapin
f
,
J.-P.
Lachaux
c
,d
,
M.
Baciu
a
,b
,
H.
Lœvenbruck
a
,b,g
a
University
Grenoble
Alpes,
LPNC,
F-38040
Grenoble,
France
b
CNRS,
LPNC,
UMR
5105,
F-38040
Grenoble,
France
c
INSERM
U1028-CNRS
UMR5292,
Brain
Dynamics
and
Cognition
Team,
Lyon
Neuroscience
Research
Center,
F-69500
Lyon-Bron,
France
d
University
Claude
Bernard,
Lyon
1,
F-69000
Lyon,
France
e
INSERM,
U836,
Grenoble
Institut
des
Neurosciences,
38700
La
Tronche,
France
f
Laboratoire
de
phonétique,
Département
de
Linguistique,
Université
du
Québec
à
Montréal,
Canada
g
GIPSA-lab,
Département
Parole
et
Cognition,
UMR
CNRS
5216,
Université
de
Grenoble,
Grenoble,
France
h
i
g
h
l
i
g
h
t
s
Behavioural,
physiological
and
cerebral
correlates
of
covert
and
overt
speech.
Inner
speech
during
reading
and
writing.
Inner
signing
and
voice-hallucinations
in
deaf
people.
Inner
speech,
mind
wandering
and
cognitive
performance.
Agency
during
inner
speech.
a
r
t
i
c
l
e
i
n
f
o
Article
history:
Received
21
August
2013
Received
in
revised
form
23
December
2013
Accepted
26
December
2013
Available online 8 January 2014
Keywords:
inner
speech
inner
voice
verbal
thoughts
silent
reading
inner
signing
inner
speech
monitoring
auditory
verbal
hallucination
verbal
mind
wandering
a
b
s
t
r
a
c
t
The
little
voice
inside
our
head,
or
inner
speech,
is
a
common
everyday
experience.
It
plays
a
central
role
in
human
consciousness
at
the
interplay
of
language
and
thought.
An
impressive
host
of
research
works
has
been
carried
out
on
inner
speech
these
last
fifty
years.
Here
we
first
describe
the
phenomenology
of
inner
speech
by
examining
five
issues:
common
behavioural
and
cerebral
correlates
with
overt
speech,
different
types
of
inner
speech
(wilful
verbal
thought
generation
and
verbal
mind
wandering),
presence
of
inner
speech
in
reading
and
in
writing,
inner
signing
and
voice-hallucinations
in
deaf
people.
Secondly,
we
review
the
role
of
inner
speech
in
cognitive
performance
(i.e.
enhancement
vs.
perturbation).
Finally,
we
consider
agency
in
inner
speech
and
how
our
inner
voice
is
known
to
be
self-generated
and
not
produced
by
someone
else.
© 2014 Elsevier B.V. All rights reserved.
Contents
1.
Introduction
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 221
2.
Definition
of
inner
speech
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 221
3.
Phenomenology
of
inner
speech
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 221
3.1.
Relationship
between
inner
speech
and
overt
speech
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 222
3.1.1.
Behavioural
and
physiological
correlates
of
inner
speech
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 222
3.1.2.
Cerebral
correlates
of
inner
speech
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 223
Corresponding
author
at:
Laboratoire
de
Psychologie
&
Neurocognition
(CNRS
UMR
5105),
BSHM,
BP47
38040
Grenoble
Cedex
9,
France.
Tel.:
+33
4
7682
5880;
fax:
+33
4
7682
7834.
E-mail
address:
perronemarcela@gmail.com
(M.
Perrone-Bertolotti).
0166-4328/$
see
front
matter ©
2014 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.bbr.2013.12.034
M.
Perrone-Bertolotti
et
al.
/
Behavioural
Brain
Research
261 (2014) 220–
239 221
3.2.
Wilful
inner
speech
vs.
verbal
mind
wandering
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 227
3.3.
Inner
speech
and
reading
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 227
3.4.
Inner
speech
and
writing
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 229
3.5.
Modalities
of
inner
speech:
inner
signing
and
voice-hallucinations
in
deaf
people
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 229
3.6.
Summary
so
far
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 230
4.
The
influence
of
inner
speech
on
cognitive
performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 230
4.1.
Inner
speech
and
task-switching
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 230
4.2.
Inner
speech
and
problem-solving
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 231
4.3.
Inner
speech
and
rumination
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 231
5.
Agency
during
inner
speech
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 231
5.1.
Agency
within
predictive
modelling
of
inner
speech
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 231
5.2.
Evidence
for
“inner
speech
is
an
action”
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 233
5.3.
A
case
of
impaired
inner
speech
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 233
6.
Conclusion.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 234
Acknowledgements
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 236
References
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 236
1.
Introduction
We
all
hear
a
voice
inside
our
brain,
commonly
called
“inner
voice”,
“inner
speech”
or
referred
to
as
“verbal
thoughts”.
Inner
speech
is
directed
by
oneself
to
oneself,
and
is
produced
in
one’s
mind.
Inner
(silent)
rehearsal
plays
a
central
role
in
the
temporary
storage
of
information
in
short-term
memory
(phonological
loop).
Silently
verbalizing
is
also
crucial
in
thinking
and
self-awareness.
It
helps
in
planning,
problem-solving,
self-motivating,
reading,
writing,
calculating
and
autobiographical
memory.
Inner
speech
can
thus
have
a
positive
influence
on
many
cognitive
tasks.
But
it
can
be
disrupted
and
have
a
negative
role.
Excessive
negative
self-reflecting
or
mental
rumination
can
interfere
with
cognitive
performance
and
has
been
shown
to
be
a
risk
factor
for
depression
and
anxiety
disorders.
It
is
also
suggested
that
inner
speech
mon-
itoring
deficit
may
result
in
auditory
verbal
hallucinations
(AVH),
where
one’s
own
verbal
thoughts
come
to
be
perceived
as
external
voices.
Before
we
examine
several
questions
related
to
the
pheno-
menology
of
inner
speech,
its
role
in
cognitive
performance
and
its
relation
to
self-monitoring,
we
will
first
define
inner
speech,
based
on
several
descriptions
from
the
literature.
2.
Definition
of
inner
speech
Inner
speech,
silent-,
imagined-,
covert-speech,
verbal
thought,
are
some
of
the
terms
used
to
refer
to
the
silent
production
of
words
in
one’s
mind,
or
the
activity
of
talking
to
oneself
in
silence
[1].
It
has
also
been
defined
as
the
silent
expression
of
conscious
thought
to
oneself
in
a
coherent
linguistic
form
[2].
Inner
speech
plays
a
central
role
in
human
consciousness
at
the
interplay
of
language
and
thought
[3,4].
It
is
a
pervasive
and
significant
human
mental
activity.
It
is
estimated
that
at
least
one
quarter
of
people’s
con-
scious
waking
life
consists
of
inner
speech
[5,6].
Inner
speech
can
be
considered
as
the
mental
simulation
of
speech.
As
such
it
can
contribute
to
the
understanding
of
speech
representations
or
of
speech
planning
and
preparation.
Inner
speech
is
central
in
various
cognitive
functions.
It
has
been
suggested
that
inner
speech
may
interact
with
working
memory
in
order
to
enhance
the
encoding
of
new
material
[7,8].
Inner
speech
is
also
involved
in
rehearsing
past
communication,
situation
and
emotions
or
in
past
situation
recall
(autobiographical
memories)
[4].
Inner
speech
also
plays
a
crucial
role
in
future
situation
planning,
in
thinking
[9]
and
in
conscious-
ness,
self-awareness
and
self-regulation
(see
[4,10,11]).
It
is
also
implicated
in
reading
and
writing
[12–14].
From
a
developmen-
tal
point
of
view,
Vygotsky
[2]
believed
that
inner
speech
derived
from
external
speech
according
to
a
gradual
developmental
pro-
cess
of
internalization,
with
younger
children
only
able
to
“think
out
loud”
(see
also
[15–17]).
According
to
Vygotsky
and
Conrad,
it
is
not
until
about
age
5
years
or
even
later
that
children’s
overt
ver-
bal
production
can
become
fully
internalized.
This
strong
claim
is
contradicted
by
recent
evidence,
however,
suggesting
that
infants
may
use
inner
speech.
Mani
and
Plunkett
[18]
have
shown
that
18-month
old
infants
can
implicitly
name
objects.
These
inter-
nally
generated
names
can
prime
infants’
subsequent
responses
in
an
object
spoken
word
matching
task.
Ngon
and
Peperkamp
[19]
have
shown
that
20-month
olds
infants
can
covertly
produce
word-forms
that
they
do
not
yet
produce
overtly,
and
can
catego-
rize
them
as
mono-
vs.
trisyllabic
items.
Therefore
an
alternative
view
is
that
children
can
produce
a
form
of
inner
speech,
a
capacity
that
could
in
fact
facilitate
their
oral
language
development.
From
a
pathological
point
of
view,
it
is
suggested
that
dysfunction
of
inner
speech
represent
symptoms
of
several
mental
health
disorders
(mental
rumination
in
depression
or
auditory
verbal
hallucinations
in
schizophrenia),
which
reinforces
its
role
in
self-awareness
and
self-regulation.
Agnati
et
al.
[20]
have
gone
a
step
further.
They
argue
that
inner
speech
is
usually
a
positive
aid
to
learning
and
reasoning.
They
propose
that
inner
speech
conferred
an
important
ability
for
the
planning
of
behaviour
by
individuals,
within
the
com-
plex
environment
typical
of
early
human
societies.
They
thus
view
inner
speech
as
an
exaptation,
in
Gould
and
Vrba’s
[21]
sense.
They
add
that,
in
certain
circumstances,
inner
speech
can
be
regarded
as
a
mis-exaptation.
When
inner
speech
becomes
too
intense,
it
can
become
a
pathological
symptom,
such
as
the
insistent
imper-
ative
inner
voices
that
characterize
schizophrenia
or
the
excessive
rumination
that
accompanies
social
anxiety
and
depression.
Over-
all,
inner
speech
is
pervasive
and
very
common
and
it
is
hard
to
prevent
oneself
from
generating
verbal
thoughts.
Yet
we
do
not
have
a
thorough
understanding
of
it
and
many
questions
can
be
raised.
3.
Phenomenology
of
inner
speech
The
nature
of
inner
speech,
its
characteristics
and
manifesta-
tions
have
been
the
focus
on
many
research
studies.
First,
the
relationship
between
inner
and
overt
speech
is
still
debated.
Are
the
behavioural,
physiological
and
cerebral
correlates
of
inner
speech
similar
to
those
of
overt
speech?
Precisely,
does
inner
speech
involve
articulatory
representations
and
are
the
cerebral
networks
involved
in
overt
speech
also
recruited
in
inner
speech?
Can
inner
speech
be
conceived
of
as
an
action
or
is
it
a
processing
of
remem-
bered
speech
percepts?
Secondly,
inner
speech
can
have
several
manifestations.
It
can
be
wilful
when
we
mentally
recite
a
poem,
but
it
can
also
be
more
unconstrained,
when
our
“mind
wan-
ders”
during
resting
states.
Is
verbal
mind
wandering
a
different
phenomenon
from
the
more
voluntary
inner
speech?
Are
there
different
cerebral
networks
associated
with
the
different
types
of
222 M.
Perrone-Bertolotti
et
al.
/
Behavioural
Brain
Research
261 (2014) 220–
239
inner
speech?
Thirdly,
one
interesting
situation
is
silent
reading.
Do
we
always
engage
in
inner
speech
when
we
read
silently?
More-
over
do
we
systematically
use
inner
speech
when
we
write?
Finally
we
will
examine
what
inner
verbal
production
consists
of
in
deaf
signers
and
how
voice-hallucinations
in
deaf
people
can
be
inter-
preted.
3.1.
Relationship
between
inner
speech
and
overt
speech
The
relation
of
inner
to
overt
speech
is
still
unclear,
and
it
is
a
highly
debated
issue
to
which
degree
they
are
related.
Inner
speech
is
generally
thought
of
as
a
truncated
overt
speech.
But
where
the
truncation
exactly
lies
is
the
matter
of
many
debates.
3.1.1.
Behavioural
and
physiological
correlates
of
inner
speech
Early
descriptions
of
the
relationship
between
inner
and
overt
speech
can
be
traced
back
to
the
behaviourist
view
of
verbal
thought.
Watson
[22]
viewed
thought
as
associated
with
covert
speech
production:
“thought
processes
are
really
motor
habits
in
the
larynx”
(p.
174).
He
further
described
covert
speech
(referred
to
as
“implicit
language”)
as
a
weakened
form
of
overt
speech,
involving
faint
actions
that
are
“abbreviated,
short-circuited
and
economized”
(Watson
[23],
p.
327).
In
other
words,
it
was
sug-
gested
by
Watson
that
movements
of
the
speech
articulators
occur
during
covert
speech.
This
strong
view
of
the
resemblance
of
inner
speech
with
overt
speech
and
of
the
presence
of
movement
during
thinking
has
been
refuted,
however.
Smith
et
al.
[24]
administered
curare
to
a
participant
to
induce
head,
lip,
jaw,
and
eyelid
paral-
ysis.
While
the
participant
was
totally
incapable
of
articulation
he
still
remained
able
to
comprehend
and
answer
questions
accurately
through
left
eyebrow
contraction.
This
experiment
suggested
that
thought,
and
inner
speech,
can
occur
without
articulation.
Later
subjective
accounts
described
inner
speech
as
frequently
resem-
bling
overt
speech,
in
that
it
is
sound-based
and
it
can
vary
in
tempo,
pitch
and
rhythm
(see
e.g.
[25]).
This
close
resemblance
has
led
some
researchers
to
argue
that
inner
speech
production
is
the
same
thing
as
overt
speech
production,
except
that
motor
execution
is
blocked
and
no
sound
is
produced.
This
“motor
simu-
lation”
view
implies
that
inner
speech
and
overt
speech
represent
a
continuum,
share
common
mechanisms
and
have
similar
physi-
ological
correlates
(see
[26],
on
covert
actions
in
general).
Among
potentially
similar
physiological
correlates
are
respiration
rate
and
speaking
rate.
As
concerns
respiration
rate,
Conrad
and
Schönle
[27]
examined
breathing
in
overt
speech,
subvocal
speech
(articulatory
movements
without
sound
production),
and
inner
speech
(with-
out
articulation
or
vocalization)
and
rest.
They
found
that
resting
respiration
changed
to
speech
respiration
along
a
continuum
from
inner,
through
subvocal,
to
overt
speech.
This
transition
manifested
itself
in
a
progressive
increase
in
the
duration
of
the
expiratory
phase.
Conrad
and
Schönle
concluded
that
this
progressive
change
towards
speech
respiratory
patterns
indicated
a
progressive
acti-
vation
of
the
speech
motor
system.
With
regards
to
speaking
rate,
Netsell
et
al.
[28]
have
examined
covert
and
overt
sentence
pro-
duction
in
persons
who
stutter
(PWS)
as
well
as
typical
speakers.
Participants
themselves
timed
their
sentences
with
a
start/stop
key
on
a
keyboard.
The
duration
of
a
sentence
was
divided
by
the
num-
ber
of
syllables
in
the
sentence.
Netsell
et
al.
found
that
speaking
rates
of
overt
and
covert
speech
were
not
significantly
different
for
typical
speakers.
But
PWS
were
significantly
slower
in
overt
speech
compared
to
inner
speech.
In
addition,
the
PWS
filled
out
a
ques-
tionnaire
regarding
the
history
and
nature
of
their
stuttering.
Six
of
the
seven
PWS
reported
they
were
100%
fluent
with
inner
speech.
The
self-reported
most
severe
stutterer
said
he
was
100%
fluent
with
inner
speech.
Altogether,
these
findings
suggest
that
inner
speech
and
covert
speech
have
similar
physiological
correlates
in
typical
speakers.
The
fact
that
stutterers
are
fluent
when
using
inner
speech
suggests
that
inner
speech
may
use
more
abstract
repre-
sentations
than
overt
speech.
This
calls
for
a
revision
of
the
motor
simulation
view.
An
alternative
hypothesis
to
the
motor
simulation
view
states
that
mental
simulation
of
speech
is
limited
to
rehearsing
early
stages
of
speech
production,
well
ahead
of
articulation
(see
e.g.
[16]).
This
latter
hypothesis
is
referred
to
as
the
“abstraction
view”,
for
which
the
experience
of
inner
speech
is
unconcerned
with
actual
motor
simulations.
It
is
opposed
to
the
“motor
sim-
ulation
view”
for
which
inner
speech
does
include
articulatory
detail.
To
further
assess
the
relationship
of
inner
speech
to
overt
speech
behaviourally,
authors
often
evaluate
speech
errors
(such
as
phono-
logical
or
phonemic
errors).
If
speech
errors
are
present
during
covert
speech
production,
like
they
are
during
overt
speech,
it
can
be
inferred
that
inner
speech
and
overt
speech
are
similar.
For
instance,
Oppenheim
and
Dell
[29]
looked
for
inner
speech
slips
via
a
tongue-twister
recitation
task.
They
specifically
focused
on
lexical
bias
(i.e.
phonological
errors
demonstrating
a
tendency
to
create
words
more
often
than
nonwords
during
tongue-twister
repetition)
and
on
phonemic
similarity
effects
(i.e.
the
tendency
for
phoneme
substitution
errors
to
occur
between
phonemes
that
share
similar?
features).
This
allowed
them
to
evaluate,
on
the
one
hand,
activation
spreading
between
lexical
and
phonological
rep-
resentations,
and
on
the
other
hand,
activation
of
sub-phonemic
feature
representations.
Their
results
suggest
that
a
lexical
bias
was
observed
in
both
inner
and
overt
speech,
but
that
the
phone-
mic
similarity
effect
was
only
observed
in
overt
speech,
suggesting
a
lack
of
phonemic
representation
(i.e.
articulatory
or
phonetic
features
representation)
during
inner
speech.
These
findings
first
suggest
that
inner
speech
has
much
in
common
with
overt
speech,
since
it
displays
slips
that
exhibit
one
of
the
most
important
error
effects
during
overt
speech,
lexical
bias.
But
they
also
suggest
that
inner
speech
is
impoverished
compared
to
overt
speech
and
that
inner
speech
is
abstract
and
underspecified
at
the
featural
level.
According
to
Oppenheim
and
Dell
[29]
there
would
be
no
feedback
of
activation
from
feature
to
phoneme
levels
of
represen-
tation,
and
thus
no
bottom-up
activation
of
competitor
phonemes.
More
specifically,
the
authors
propose
that
either
poor
generation
of
features
during
inner
speech
production
or
poor
sensitivity
to
features
during
its
perception
could
suppress
the
effect
of
phone-
mic
similarity.
To
further
test
the
hypothesis
that
inner
speech
is
abstract,
Oppenheim
and
Dell
[30]
contrasted
two
forms
of
inner
speech,
standard
inner
speech
without
articulatory
movement
and
an
intermediate
form
with
silent
articulation.
They
showed
that
the
intermediate
inner
speech
elicited
phonemic
similarity
effects,
just
as
overt
speech,
while
the
standard
inner
speech
did
not.
They
interpretation
is
that
silent
articulation
increases
the
amount
of
activation
of
articulatory
(subphonemic)
features,
thereby
increas-
ing
the
number
of
errors
involving
similarly
articulated
phonemes.
In
contrast,
they
observed
a
lack
of
phonemic
similarity
effect
in
standard
unarticulated
inner
speech.
According
to
them,
these
findings
support
a
modified
version
of
the
abstraction
view,
the
“flexible
abstraction
hypothesis”:
Inner
speech
has
multiple
facets,
it
may
fail
to
involve
articulatory
representations
but
it
can
incor-
porate
lower
level
articulatory
planning
when
speakers
silently
articulate.
A
somewhat
different
view
is
taken
by
Brocklehurst
and
Corley
[31].
In
a
similar
experimental
protocol
as
that
used
in
Oppenheim
and
Dell
[29],
they
observed
a
phonemic
similarity
effect
during
inner
speech.
Brocklehurst
and
Corley
[31]
propose
that
the
lack
of
phonemic
similarity
effect
reported
in
Oppenheim
and
Dell’s
study
[29]
may
be
related
to
the
fact
that
such
errors
may
be
difficult
to
perceive
internally.
By
adding
an
auditorily
masked
condition
and
by
using
a
mix
of
real-
and
non-words,
they
found
phonemic
similarity
effects
of
similar
magnitudes
in
both
inner
and
overt
M.
Perrone-Bertolotti
et
al.
/
Behavioural
Brain
Research
261 (2014) 220–
239 223
speech.
This
suggests
that
inner
speech
is
not
impoverished
at
the
featural
level
as
previously
claimed.
Furthermore,
in
a
later
study
[32],
using
the
real
words
material
from
Oppenheim
and
Dell
[30],
they
found
no
interaction
of
phonemic
similarity
with
overtness
in
predicting
the
likelihood
of
a
substitution
error.
The
effects
of
phonemic
similarity
were
manifest
in
both
inner
and
overt
speech,
both
with
and
without
auditory
feedback.
Their
interpretation
is
that
inner
speech
must
be
specified
at
the
sub-phonemic
level
and
that
inner
speech
is
produced
in
much
the
same
way
as
overt
speech.
The
fact
that
different
results
were
obtained
by
different
research
groups
could
suggest
that
inner
speech
in
fact
consists
of
different
subtypes
or
levels
and
that
in
each
of
the
subtypes,
flexibility
is
at
play,
as
argued
by
Oppenheim
and
Dell
themselves.
In
that
vein,
Geva
and
colleagues
distinguish
two
levels
of
inner
speech,
an
abstract
and
a
concrete
one
[33].
They
define
abstract
inner
speech
as
the
‘language
of
the
mind’,
which
involves
the
activation
of
abstract
linguistic
representation.
According
to
them,
abstract
inner
speech
has
been
the
object
of
many
research
works
in
experimental
psychology,
from
Egger
[34]
and
Ballet
[35]
and
including
Vygotsky
[2],
to
Dell
and
Repka
[36],
for
instance.
The
second
level
of
inner
speech,
or
concrete
inner
speech,
is
charac-
terized
by
Geva
et
al.
[33]
as
flexible,
being
either
phonological
or
phonetic.
The
research
works
by
Oppenheim
and
Dell
[30]
or
Brocklehurst
and
Corley
[31]
presented
above
are
related
to
this
level.
The
major
difference
between
these
two
levels
is
the
potential
involvement
of
low-level
phonetic
information
(such
as
articula-
tory
features)
during
concrete
inner
speech
only.
Different
levels
of
inner
speech
are
also
posited
by
Fernyhough
[37].
Elaborating
on
Vygotsky’s
idea
that
inner
speech
is
the
end
result
of
a
gradual
process
of
internalization
of
dialogue,
Fernyhough
has
sketched
out
a
model
of
inner
speech
development
that
includes
four
levels.
At
level
1
(external
dialogue),
children
and
caregivers
engage
in
overt
dialogue.
At
level
2
(private
speech),
children
conduct
these
dia-
logues
in
their
own
overt
private
speech,
which
becomes
gradually
subvocalised.
At
level
3
(expanded
inner
speech),
private
speech
is
fully
internalized
and
covert
but
its
dialogic
quality
is
still
man-
ifest.
At
level
4
(condensed
inner
speech),
syntactic
and
semantic
transformations
convert
inner
speech
into
an
abbreviated
dialogue.
This
last
level
corresponds
to
the
stage
of
“thinking
in
pure
mean-
ings”,
described
by
Vygotsky,
in
which
inner
speech
has
lost
most
of
the
acoustic
and
structural
qualities
of
external
speech.
Inter-
estingly,
Fernyhough
suggests
that,
even
in
adulthood,
movements
between
the
levels
may
occur.
Under
demanding
conditions,
such
as
cognitive
load
or
stress,
there
could
be
a
transition
from
fully
con-
densed
inner
speech
back
to
expanded
inner
speech
and
even
back
to
private
inner
speech.
Although
the
plausibility
of
the
gradual
developmental
process
is
debatable
(see
Section
2),
the
proposi-
tion
that
inner
speech
evolves
along
varying
levels
has
intuitive
appeal.
Who
hasn’t
“thought
out
loud”
in
conditions
of
danger
or
stress?
3.1.2.
Cerebral
correlates
of
inner
speech
Further
information
about
the
relationship
between
overt
and
covert
speech
comes
from
the
comparison
between
neural
cor-
relates
of
inner
and
overt
speech
(see
[38]
or
[39]
for
detailed
reviews).
In
agreement
with
the
view
that
inner
speech
and
overt
speech
may
belong
to
the
same
continuum,
many
studies
have
shown
common
networks
of
activation
in
tasks
that
involve
either
covert
or
overt
speech.
In
general,
both
overt
and
covert
speech
tasks
induce
activation
of
essential
language
areas
(Broca’s
and
Wernicke’s
areas,
inferior
parietal
lobule)
in
the
left
specialized
hemisphere
(e.g.
[40–46]).
However,
several
neuroimaging
studies
that
directly
include
within-subject
comparison
of
inner
and
overt
speech
tasks
[33,40–42,44,47–55],
suggest
that
although
inner
and
overt
speech
share
a
common
cerebral
network,
they
engage
some
cerebral
regions
in
different
ways
and
they
produce
separate
acti-
vations
in
other
cerebral
regions
(see
Table
1).
Some
studies
have
found
greater
activation
in
overt
speech
rel-
ative
to
inner
speech
in
motor
and
premotor
cerebral
regions
(e.g.
[44]).
These
findings
are
consistent
with
the
continuum
perspec-
tive,
in
which
inner
speech
is
considered
as
covert
speech
with
a
blocking
of
execution
processes.
Other
studies
suggest
that
covert
speech
is
not
just
inner
speech
with
added
motor
processes,
how-
ever.
Greater
activation
has
been
observed
in
overt
relative
to
inner
speech
tasks
in
regions
such
as
the
left
inferior
frontal
gyrus,
the
left
insula,
sensory
areas,
the
left
superior
temporal
sulcus,
the
supramarginal
gyrus,
the
anterior
cingulate
or
the
left
and
right
paravermal
lobule
VI
(e.g.
[40,42,44,48,50,51]).
Greater
activation
in
auditory
areas
has
been
postulated
to
be
associated
with
the
per-
ception
of
one’s
own
speech
and
greater
response
of
the
left
insula
has
been
related
to
its
role
in
articulatory
planning.
Conversely,
greater
activation
was
observed
in
inner
speech
relative
to
overt
speech
in
several
regions
including
the
left
precen-
tral
gyrus,
left
middle
frontal
gurys,
left
or
right
middle
temporal
gyrus,
left
superior
frontal
gyrus,
right
cingulate
gyrus,
left
or
right
inferior
parietal
lobe,
left
dorsal
frontal
cortex,
left
parahippocam-
pal
gyrus,
right
cerebellum
(e.g.
[41,48,51]).
These
observations
imply
that
covert
speech
generation
cannot
be
simply
equated
to
overt
responses
minus
articulatory
motor
execution.
Several
inter-
pretations
have
been
proposed.
First,
some
of
the
authors
cited
above
have
suggested
that
increased
activity
in
inner
compared
to
overt
speech
could
reflect
greater
overall
signal
variance
in
overt
conditions
due
to
greater
head
motion.
Secondly,
some
of
these
activations
(e.g.
in
inferior
and
superior
parietal,
superior
frontal,
anterior
cingulate,
and
middle
temporal
regions)
have
been
attributed
to
inhibition
of
overt
responses
or
movement-related
conflict
(producing
a
word
but
not
saying
it
aloud).
The
varying
results
can
be
explained
by
the
different
types
of
inner
speech
tasks
involved
in
the
studies
listed
here.
The
different
tasks
(which
include
word
repetition,
letter
or
object
naming,
animal
name
generation,
verb
generation,
reading,
rhyme
judge-
ment,
counting)
involve
different
speech
production
processes,
from
lexical
retrieval
to
phonological
or
even
phonetic
encoding.
These
processes
engage
different
brain
regions
and
are
associated
with
different
degrees
of
monitoring,
i.e.
with
varying
degrees
of
awareness
(see
e.g.
[56]).
Rhyme
and
homophone
judgement
tasks,
for
instance,
involve
active
use
of
inner
speech.
In
these
tasks,
participants
have
to
monitor
their
own
inner
speech,
keep
the
first
presented
item
in
working
memory,
while
they
process
the
second
item.
Phonological
awareness
is
crucial
in
these
tasks,
as
well
as
a
high
level
of
attention
to
one’s
inner
speech.
In
seman-
tic
fluency
tasks
(such
as
word
generation),
monitoring
is
also
crucially
involved,
to
keep
track
of
the
words
already
produced.
In
contrast
tasks
such
as
word
repetition,
recitation,
letter
naming,
silent
reading
or
counting,
require
less
monitoring.
Evidence
for
the
fact
that
different
degrees
of
awareness
to
one’s
own
inner
speech
lead
to
different
patterns
of
brain
activations
was
provided
by
an
fMRI
study
of
the
verbal
transformation
effect
[57].
‘Verbal
transformation’
refers
to
the
perceptual
phenomenon
in
which
listeners
or
speakers
report
hearing
a
new
percept
when
a
word
is
repeated
rapidly
[58].
In
the
study
by
Sato
and
colleagues,
partic-
ipants
were
asked
to
silently
repeat
pseudo-words
such
as/psə/.
In
the
baseline
condition,
participants
were
asked
to
covertly
repeat
the
sequence
over
and
over
(/psə
psə
psə
psə
psə
.
.
./).
In
the
verbal
transformation
condition,
they
had
to
covertly
repeat
the
sequence
and
actively
search
for
a
transformation
(from/psə/
to
/səp/
for
instance).
When
compared
with
the
baseline
condition,
inner
speech
with
active
search
for
verbal
transformation
correlated
with
stronger
activation
in
the
left
inferior
frontal
gyrus
and
left
supramarginal
gyrus,
as
well
as
other
regions
(anterior
part
of
the
right
cingulate
gyrus,
bilateral
cerebellum
and
left
superior
224 M.
Perrone-Bertolotti
et
al.
/
Behavioural
Brain
Research
261 (2014) 220–
239
Table
1
Brain
activation
ˆ
patterns
observed
in
studies
that
have
directly
compared
overt
vs.
covert/inner
modes
in