Graduate Seminar: Interpretation of Libet Experiments

Seminar prepared for UCL Department of Philosophy, 2011. Commentaries on Libet (Brain 1983), Mele, Libet 1985, Haggard & Libet (2001), Levy (JCS 2005), Hacker (PFN section 8.2), Gomes (C&C 1998), Churchland (1981), Schlosser, Kane 1998, Searle, Davidson, Cairney 1975, and Haggard (NRN 2008).
The lecture handout is available as pdf.

Libet - Brain 1983

electrical potential at the brain’s surface

proportional to a summation of many thousands of neurones firing synchronously.


“get ready” tone → 1-3s fixation → sweeping clock, 2.56s revolution

Clock visible ticks spaced every 107ms

→ wait 1 rev, then “at any time thereafter when you feel like doing so”, press

Electrical activity of muscle measured – movement has to be brisk, abrupt

→ continuation interval 500-800ms (discrete to 15ms)

→ “within a few seconds” report as they remember it after the event. (to prevent worrying about the task in advance)

  • ExpWA: sic – report absolute time A on clock at moment of first “wanting to perform” movement. Ss described as “urge, intention, decision to move”. Ss reported, when specifically asked, that a minority of acts occurred “without a conscious prior intention” – acts “surprised” them. Ss easily distinguished W form preplanning, and urge not followed by actual movement.
  • ExpWO: after continuation, clock jumps to position in a 600ms window (discrete to 15ms) around the previously reported WA time. Report if spot earlier or later than want/urge to move.
  • ExpMA: awareness that they actually moved. “mental set differed in W vs M blocks”: “feeling of active attention to (or watching for) the awareness of wanting to move”.
  • ExpMO: simile with Order

ExpSA: tactile near-threshold stimuli between 2.6-7.6s. Report “time of awareness of the sensation”. In order to capture bias in “making simultaneous judgements of the times for a mental/sensory event and a visual event”.

2 subjects: “let the urge to act appear on its own at any time without any preplanning or concentration on when to act” (capricious, spontaneous)


RPs: Type I: -700ms. preplanning was reported

Type II: -500ms, spontaneous/capricious instructions

Type III: -250ms


Averaged 40 trials, and averaged W for same block.:

(W – EMG) = -204ms, (M – EMG) = -86ms, (S – stim) = -47ms

(RP – EMG) = -550ms; (RP – W) = -350ms


  1. “It is clear that neuronal processes that precede a self-initiated voluntary action, as reflected in the readiness-potential, generally begin substantially before the reported appearance of conscious intention to perform that specific act.”
  2. “some neuronal activity associated with the eventual performance of the act has started well before any recallable conscious initiation or intervention could be possible.”
  3. “the brain evidently ‘decides’ to initiate, or at least, prepare to initiate the act at a time before there is any reportable subjective awareness that such a decision has taken place. It is concluded that cerebral initiation even of a spontaneous voluntary act... can and usually does begin unconsciously” [my italics].

Premise: subjective event in question is only introspectively accessible to the S himself, and this requires a report by S.

Consideration 1: simultaneity of urge and clock position. Ss “observe simultaneously, for later report, the appearance of a mental event” & the clock. Prior entry for attending is ‘compensated’ by the S task. [is it? Who knows what subjects are attending to when they judge W.]

Consideration 2: report of awareness is “the primary evidence for introspective experience”, but we can conceive/postulate conditions that might introduce discrepancies with actual-vs-reported times of such awareness:

  • it is possible only the end of the mental event is accurately judgeable
    • though subjectively no difficulty was raised
    • and Efron 1973 found perceptual onset latency is constant wrt length of stim
    • and perceptual offset latency varies when stim duration < 150ms
  • ‘preawareness’ that one is preparing to perform the voluntary act “would mean that the reported W times were earlier than they should have been” and W-RP is even more negative.
  • Possible earlier nonrecallable phase of conscious urge – not stored in STM.
    but Ss don’t need to store event itself; just clock position (S controls for this)
    [so, possible earlier non-bindable-to-vision phase of conscious urge?]
    and would be hypothetical and not testable.

Consideration 3:

Reasons for believing valid

  • 2 converging methods
  • subjective and objective ability to distinguish W from M
  • similarity of RP in different blocks

but W is 50ms earlier if subjects performed W judgements after having performed M.

[subjects trying to please the experimenter]

[unnatural scenario]


[More difficult to maintain:] “These considerations would appear to introduce certain constraints on the potential of the individual for exerting conscious initiation and control over his voluntary acts”

  • Conscious veto – possible because W is still about 150-200ms before EMG. [and 100ms before M?]. Subjectively ‘aborted’ urges.
  • Deliberation is different


Alfred R. Mele: Free will: Action theory meets neuroscience

  1. people “do not need to have an opinion about whether determinism [that laws of nature & complete description at t1 logically entails a complete description at t2] is true” to use the term ‘acting freely’. But this does not imply free will and determinism are in general incompatible
  2. Frankfurt 1988: when people act intentionally without compulsions, psychosis, addiction, duress, they act freely. An act’s being caused deterministically is insufficient for its being coerced. Not “S could not have done otherwise”, but “if the situation had been different in specific ways, S would have done otherwise” – which is compatible with determinism
  3. Libertarian viewpoint (Kane): people do act freely, and this is incompatible with determinism. ‘Decidings’ are mental events, and obey probabilistic laws.
  4. Deciding to A “is to perform a momentary mental action of forming an intention to A”
    1. Feeling an urge is not deciding because “you can have an urge to A but decide not to”
    2. Urges and wants can be directly opposing, but intentions cannot. “Intending to A while intending to refrain from A, if possible at all, would be a sign of a serious disorder”
    3. Deciding is not a process but a momentary action
    4. “Not all intentions are formed in acts of deciding” e.g. routines
    5. “All decisions are prompted partly by uncertainty about what to do”
  5. Intensions and decisions can be
    1. Distal or proximal (in time)
    2. Relatively specific or unspecific (in relation to possible acts)

Conceptual difference between motivational states (desires & urges) and executive states (choices & formations of intentions). Libet conflates the two.


Libet 1985:

Vetoing experiment – Ss plan to press at a given time, but veto it. RP alters in direction at 150-250ms before the preset time – i.e. roughly the time that W appears.

Veto RP vs preplanned RP vs capricious RP

Preplanned RPs are same as those of subjects who are neither told to flex at a preset time nor instructed to flex spontaneously, and as those who report some preplanning in the spontaneous task (type I). capricious = type II

Veto RPs are type I until 150-250ms before preset time.

Mele: what is vetoed is not an “intended motor action”, because one cannot “intend to A” and “not intend to A” simultaneously. Perhaps they have an urge.


  • All RPs represent proximal urges; some earlier than others
  • Final phase of RP corresponds to proximal intention – absent in veto
  • Proximal urge (urge to act now) → proximal intention
  • Urge to perform A anytime soon → proximal intention
  • Urge to perform A at a specific time in the near future → proximal intention

Mele believes no proximal intention is present at RP onset, and RP cannot represent proximal intention. Proximal intention is apparently formed between RP onset and action, and “for all Libet’s data show, those intentions may be consciously formed or acquired”.


Haggard & Libet 2001:

“an unconscious gap of ~400ms between onset of the cerebral process and when the person becomes consciously aware of the decision or wish or intention to act. If the ‘act now’ process is initiated unconsciously, then conscious free will is not doing it” [my italics]. Mele believes

  • the time of ‘act now’ is like when you start baking a pizza.
  • “Who would have thought that conscious free will has the job of producing urges?” – the early part is not a decision
  • Processes have parts: the ‘act now’ may have an early, distal, unconsciously initiated component – and a later, more proximal, consciously initiated component.
  • “that urges to do things arise unconsciously – urges on which the agent may or may not act about half a second after they arise – is no cause for worry about free will”

Intentions can form as close as 230ms before act (e.g. simple RT), if there is no decision (no uncertainty). Stimulus + Conditional intention to act on stimulus → intention. Alternately, there may be no proximal intention, and Stim + conditional intention → action. So in Libet, maybe urge + conditional intention → action

If intention is at work, then it’s “a good bet they are acquired around time W”.


Neil Levy (Oxford Centre for Neuroethics), JCS 12 p67, 2005

Libet’s impossible demand

  • Libet says though we do not initiate actions consciously, we might be able to consciously veto them. But Clark 1999: “If unconscious RP must precede initiation of action, why does it not equally need to precede the vetoing of an action?”
  • Libet argues that if we do not exercise conscious control, we do not exercise free will.
  • “[Libet’s] challenge to free will is a challenge to compatibilists and libertarians alike” – “he is claiming that, in some sense, we do not originate our actions at all”

Arguments against Libet

  • Flanagan 1996: ‘big picture’ decisions may be consciously initiated, while details may be left to subpersonal processes. → responsibility tracks back to Ss “setting themselves” to respond as asked
  • Mele 2005: Unconscious events tracked by RP are better identified with urge/desire, not intention/decision. → Libet has not shown that our decisions are not consciously made.
  • Haggard + Libet JCS2001: “Though we do not consciously initiate actions, conscious intention may coincide with the specification of action.” (LRP)→ there remains a role for free will.

All leave this (weak) statement uncontested: “That, if we are not conscious of at least some of our decisions to initiate an action at the very moment of initiating it, we do not exercise free will in initiating it”. But:

  • Libet asserts unconsciously developed actions cannot be regarded as free will. I.e. decision-making must be conscious if it is to be free.
  • Rosenthal: volitions need not be consciously caused in order for them to count as free. (e.g. as long as they fit with our conscious self-image). But
    • My act “not fitting with my self-image” can’t defend against being responsible for the act
    • Simply accepting I have Tourettes does not make me responsible.
  • Dennet 2003 defence: self should not be identified with brain parts, or the conscious subsystem; rather, “the entire set of cognitive processes, personal and subpersonal”. “You are not out of the loop; you are the loop”.
  • But
    • passing thoughts and sudden urges don’t fit. Utilisation behaviour, Tourettes, states of automatism/disinhibition } not considered as free.
    • Freeman 1999: It is valuable for me to hold myself responsible for all my actions, but it seems unjust for others to hold me responsible for actions over which I had no control

Libet’s demand is impossible to fulfil: requirement that we need to consciously initiate our actions, to be free.

Levy argues we can deduce that, if conscious decision-making is a condition for free will, we can’t have it.:

  • Phenomenology: Dennet “We do not witness [our decision] being made; we witness its arrival
  • Exclusion of what conscious deliberation could be like. If deliberation occurs consciously (eg choosing a job), the information upon which deliberation operates (weightings of favouring particular reasons) must come either from either
    • subpersonal processes, independently of deliberation. We consciously ‘recognise’ the subpersonally generated weights.
    • higher-level reasons – but then these reasons need weights too, which must be independent of conscious deliberation.
    • if they are neither reasoned nor unconscious, they could only be arbitrary
  • → “If there were conscious will, it would not be free will; it would be mere random chance”
  • Infinite regress: “the demand that we exercise conscious will seems to be the demand that we control our controlling. And that demand cannot be fulfilled.”
    • the actual decision is not made consciously “the conclusion that the first or second set of reasons is weightier – is simply reported to consciousness”
    • Saying we can decide to veto these is subject to same issue
    • Conscious deliberation can help to select motivations: deliberation as ‘realising’, ‘comprehending’. But then, I only control it “inasmuch as I can cease to engage/persist in it”

→ Our failure or success at being conscious of our decisions as we make them is irrelevant to our freedom, since those decisions must be the product of unconscious mechanisms in any case.

Peter Hacker PFN 8.2: Libet’s theory of voluntary movement

Presuppositions of experiment:

  1. ‘subjective event is only accessible introspectively to the subject himself’
  2. ‘Each subject was instructed to “watch for” and report the earliest appearance of the awareness in question’


  • “Saying that someone did something because they wanted to is not to give a causal explanation”

  • William James is confused: ‘willed action’ as ‘action performed when we consciously pay attention to its selection’
  1. It is neither necessary nor sufficient for an act to be voluntary that it be preceded by a feeling of want/urge/intention
    1. not necessary: e.g.
      1. picking up a pen to write is voluntary
      2. going along with earlier preplanned intentions
    2. not sufficient: e.g. sneezing is involuntary (can be inhibited but not caused by volition)
  2. “Libet’s theory would in effect assimilate all human voluntary action to the status of inhibited sneezes or sneezes which one did not choose to inhibit”
  3. “a movement that is caused by an urge is precisely not a voluntary action” – e.g. I phi’d because I had the urge to: the ‘because’ is not causal

→ finding is merely that, people can find feelings of intention/urge to move, when they are asked to, though such things don’t actually occur normally. “The feeling reported is not what makes their movement voluntary”


Gilberto Gomes (Brasil) Consc&Cog 1998: The timing of conscious experience: a critical review and reinterpretation of Libet’s research

Agrees that RP is “a consequence of neural processes that correspond to mental events” (may be subconscious) → “one would be inclined to consider it as the cause [of a movement]”

“but the cause of a voluntary movement is usually considered to be a conscious decision” - what is the relation between decision and awareness?

Are they really acting freely? “That the moment of the act should not be predetermined” is a predetermined condition

Attacks experimental technique

  • Divided attention + spontaneous movt: Comparisons of trials on which Ss reported being not conscious of having moved → similar RP.
    • → RP is for all spontaneous acts, whether conscious or not.
    • But Libet explains it as amnesia
    • Amplitude greater for conscious movts., and there is an extra peak at +10ms.
    • RP earlier onset for unconscious movts.
  • Suggestive force of the experimenter
    • Breitmeyer: performed the experiments “without awareness of intent to act. By requiring subjects to attend to awareness of intent, Libet may have imposed intention artificially.”
    • Scheerer 1985: “In introspection we often find those events that we have been led to expect”
      • Libet justifies with subjective reports – occasionally some Ss are “taken by surprise” as if the movement had arrived “on its own”
      • Libet says subjects had no difficulty in distinguishing W and M moments.
    • But this does not mean that the decision to act and the act itself are experienced as discrete events.
    • And some Ss felt a difference in “mental set” for W and M tasks
  • Extensive training: cites differences between subjects who did M then W, vs W then M
  • Latency → can’t be certain of times; in particular simultaneity
    • Note Eccles: “The andedating of sensory experience is attributable to the ability of the self-conscious mind to make slight temporal adjustments, i.e. to play tricks with time.” → Dualism
    • Penrose: reversal of physical time.
    • S is near threshold → the latency for awareness of S probably >> latency for perceiving the clock face.
    • Clock is a complex shape → latency for perceiving clock position >> latency for simple S
    • Libet invokes ‘subjective referral’ of event times backwards to compensate for sensory latency.
      • Gratuitous; a more parsimonious explanation is an additional cortical-activity-to-consciousness latency, which varies between modalities (and possibly stimuli)
    • “Since the reported clock-times are somewhat unprecise assessments dome by the subject, to which we should add an unknown value, we conclude that we can attach no value at all to such reports as an indication of the real time of occurrence of the conscious experience”

“conscious decision” = 2 senses:

  1. conscious because it results from conscious awareness of something
  2. conscious because one is conscious of it

Conscious (1) control functions may occur at first nonconsciously (2), after conscious awareness of intention to act, and then become conscious themselves.

  1. nonconscious decision to act
  2. consciousness of decision to act
  3. nonconscious veto (this is conscious control because it results from 2)
  4. inhibition of movement
  5. consciousness of veto

Churchland 1981 On the alleged backwards referral of experiences

Caricature of Libet’s hypotheses:

“There is an automatic subjective referral of conscious experience backwards in time…after the delayed neuronal adequacy has been achieved.”

  1. there is a temporal displacement of one sensation such that it appears earlier in the order of experienced events than it should
  2. the skin sensation is felt earlier than the brain states necessary for its production
  3. subjective time is different from physical time
  4. mental states cannot be identical to physical states


  1. the evidence that there is a temporal misordering of experiences is inadequate
  2. even if there is, this is no evidence for a non-material mind

Schlosser, Phil Psychol: Free will and the unconscious precursors of choice

Narrow objection: alternative explanation of findings e.g. RP represents urge not decision/intention

Broad objection: Libet’s argument is: final motor decision is predetermined → no free will. But:

  • Ignores distal intentions, which somehow “work their way into” the motor system, and are efficacious.
  • “When?” decisions are usually insignificant compared to “What?” decisions
    • Moral responsibility, seeing a doctor etc

Generous interpretation of Libet: RP may correspond to a process that will result in P-intention+movement unless vetoed. → P-intention is not the causal initiator as it occurs too late → consciousness merely accompanies the process causing the movement

  • Soon & Haynes 2008: fMRI decoding determines which 2-alternative ‘free choice’ selection Ss make ~10s beforehand. → predictive of What
  • Moving to a distal intention only “pushes the problem one step backward” – distal intention may also have an unconscious precursor.
    • Unconscious precursor could just indicate motor preparation?
  • Soon et al. activation is in nonmotor areas. Other studies = prediction with “I would buy this item” (hypothetical and distal intentions)

Other approach

Free will = (1) Absence of constraint & (2) Ability to do otherwise

Robert Kane 1998 adds “(3) agent is the origin or source of the choice and action”

Schlosser argues that (2) can be expressed in a form compatible with determinism, using counterfactual conditionals: “S can do otherwise ↔ S would do otherwise, if S <had chosen/wanted> to do otherwise”

Searle: “there’s no mental reality to the formation of a RP” – it does not indicate an unconscious mental phenomenon.

Pettter Johansson (UCL) – Choice blindness, face preference subverted by card trick



Donald Davidson: Freedom to Act

GE Moore: “A can phi” = “If A chooses to phi, then he phi’s”

JL Austen: attacks this with a faulty argument: “A phi’s if he chooses” translates with this rule to “A can phi if he chooses”. Therefore “A phi’s = A can phi”. (Flawed, because “A can phi if he chooses” is not a causal conditional – it entails its own consequent)

Austin then argues for 2 different ifs – one is causal conditional, one isn’t. But is this right? Consider “if you eat apple seeds you will get arsenic poisoning” (causal conditional) vs “if x is divisible by 4, x is divisible by 2” (not causal conditional). The differences is not semantic but epistemological.

Contraposition is valid for causal conditionals. So “A can phi” =“A doesn’t phi implies he didn’t choose to”

Chisholm: But can’t extend this definition to trying/choosing, because one may not be able to try/choose. In general “tha antecedent of a causal conditional that attempts to analyse ‘can’/’could’/’free’ must not start with a verb of action, i.e. a verb that makes sense in the question “can someone do x”.

Cairney Acta Psychol 39:329-40 1975

Bisensory order judgement and the prior entry hypothesis

Sternberg & Knoll 1973: prior entry of one modality when attended. Is this decision bias or channel speeding? “If attending affects only the decision mechanism, then the differences in the means of psychometric functions for order judgement should be the same for all intensities.” Converseley, if latency of one channel were reduced, then differences in means should diminish as the intensities of the two signals increase.

But: could be response bias because of instruction ‘to attend’

Expt1: range: visual 70ms before – 30ms after the auditory stimulus. Conditions:

1 – simple temporal order. 2 – easy visual discrimination task (longer of 2 lines) then temporal order. 3 – difficult visual discrimination task then temporal order. Also rate confidence of order judgement 1—6.

2arcsin(sqrt(p)) and bias=scale category point at which P(hit)+P(false alarm)=1.

Evidence supports does not support prior entry, rather decreased accuracy with attention to one or other streams.


Haggard NRN2008