The critic has a job. This week you sit down with it and ask what that job has been: what it protects, what it fears, and why it learned to speak first.
The job the critic has been doing since you were young, why fighting it makes it louder, and the protective work each of the four critic types is carrying. Watch this first, then work through the practice below.
This module asks you to sit with the critic again and listen to what it has been protecting you from. That can bring up feelings, faces, and moments you have not thought about in years. Thirty seconds of grounding first, even if you feel steady.
Let your breath follow the circle. Out as it shrinks, in as it grows.
These pages are a practice guide. They are not therapy. If the dialogue opens something older or harder than you can carry on your own, step out of the work, ground, and take it to a therapist. If you are in crisis, reach out now: your country's helpline is at findahelpline.com. Singapore: SOS 1767. Australia: Lifeline 13 11 14. UK: Samaritans 116 123. US: 988. This work is not designed for acute states: recent trauma, active suicidal thoughts, an active eating disorder, recent grief, or an unsafe relationship.
Start here. The workbook is where you write what the critic says, so open it before you begin the practice below. It starts with the practice pages you write in during the audio, then your reflection, then the teaching: the job the critic has been doing, the fears that come up when you imagine letting it go, what the research actually shows, and this week's experiment.
A guided practice in my voice. You bring the critic to mind and ask it what it has been trying to protect you from, how old it feels when it is doing this job, and what it is afraid would happen if it stopped. Have the workbook open at the practice pages before you start, and write your answers as they come.
Meeting the Protective Function · guided audio
Everything above is complete. If you stop here, you have done Module 2 in full. What follows is one more layer, for the week you want it.
This one looks further back than your own life. You picture what came down to you through the people before you, and you set it down at the point where it reached you. It asks nothing of your actual family, and you do not have to name anyone. It is a felt image rather than a technique. The PDF has who it is not for, and when to leave it for another week.
What Came Down the Generations · guided audio
What Came Down the Generations (PDF)The practice on its own pages, with space to write. Ground first, and stop at any point.Optional. If you want to know what this module is built on, the modalities, the research, the reason each move works.
For each move this module asks you to make, here is why it works: the mechanism, the therapies that use it, and what the research actually shows.
What we are actually doing. The reframe from enemy to protector ends an internal war. As long as the critic is the enemy, the system is at war with itself, and a system at war spends its energy defending rather than shifting. Once the harsh voice is read as a strategy that once made sense, the energy that went into the fight becomes available for the work. You cannot renegotiate with something you are trying to destroy.
Two consequences follow. A protector does not stand down because it loses an argument. It stands down when it registers that its job has been seen, because its whole function is to watch for a danger it believes no one else is tracking. Turning toward it with warmth recruits the soothing and affiliative system, and the soothing and threat systems inhibit each other, so warmth switches the channel the exchange runs on. Argument keeps the critic on the threat channel and confirms the danger.
The moral charge also moves. A threat-detection function running dominant for decades describes a system in a configuration, which you can rebalance, and which took that shape for reasons that made sense at the time.
Used across: IFS (there are no bad parts, every protector carries a positive intention, curiosity toward the part as the way in); CFT (self-criticism as a safety strategy of the threat system, met by building the soothing system rather than by argument, inside Gilbert's three-systems model of threat, drive and soothing); ego-state and Voice Dialogue work; polyvagal thinking maps similar territory as ventral, sympathetic and dorsal states.
Evidence. This is a clinical model rather than a single tested ingredient, and it should be held that way. IFS has one small proof-of-concept randomised trial in rheumatoid arthritis (Shadick et al., 2013), where patients reported less pain and lower mood symptoms afterwards, while the disease activity a rheumatologist measured did not shift. The authors called it a feasibility study and asked for proper efficacy trials. There is also a small pilot for depression (Haddock et al., 2017), too small to tell the approaches apart. Neither isolates the reframe or targets self-criticism. Longe et al. (2010), a small brain-imaging study, found self-criticism and self-reassurance recruiting different regions, and Depue and Morrone-Strupinsky (2005) model affiliative bonding through dopamine and opioid pathways, with oxytocin in a modulating role. Those support the threat-versus-soothing switch the move relies on. Gilbert's three-systems model is an evolutionary and neurobehavioural framework whose support is indirect and convergent, consistent with the model rather than proving it.
Why it works. A system at war with itself cannot change, and a protector eases when it feels understood rather than when it is argued with.
What we are actually doing. The instinct to fight the critic runs into two walls. The first is suppression. Trying not to think a thought reliably feeds it, because monitoring for it keeps it primed, so telling the critic to be quiet is part of what keeps it loud. The second is escalation. A protector reads a head-on attack as evidence that you do not grasp the danger it guards against, so it works harder. Counter-affirmations fail on the same logic, a second voice arguing with the alarm while the alarm keeps ringing. Both are moves against the part, and moves against a protector intensify it. The only route that does not escalate is to meet it.
Used across: CFT (the threat-and-soothing logic), ACT (acceptance and defusion rather than argument), the mindfulness-based therapies (non-suppression as a first principle), IFS (relationship with the part instead of opposition). Across acceptance-based, compassion-based and parts-based work, nobody recommends pushing the voice away.
Evidence. Wegner et al. (1987), the white-bear studies, where trying not to think of something made the thought return more often afterwards. The meta-analysis by Abramowitz et al. (2001) found a small-to-moderate rebound effect. Later work suggests it is smaller and less reliable than the original studies implied: Wang, Hagger and Chatzisarantis (2020) found the rebound held across the studies they pooled, though the effect was no longer statistically significant once small-study bias was corrected for, and the authors treat that correction as suggestive rather than conclusive. That a protector escalates when attacked is convergent clinical observation across the parts-work field rather than a trial result.
Why it works. Pushing a thought away primes its return and attacking a protector confirms the danger, so the only move that does not feed the critic is to meet it.
You re-enter a recent occasion sensorily rather than describing the critic, and let it arrive the way it actually arrived. Emotional material comes back more readily when the state it was encoded in is reinstated (Bower, 1981; Eich, 1995), though the effects vary by design and show most reliably when the remembered material is internally generated rather than presented from outside, and an emotional learning has to be live for anything to reach it (Foa and Kozak, 1986; Ecker, Ticic and Hulley, 2012, whose reconsolidation account rests on solid animal work, while the human evidence is disputed, with the best-known human study since corrected and its replications mixed. Treat it as a reasonable clinical model rather than a settled finding). Describing the critic gets you your already-tidied account of it. Two situations rather than one, because what repeats across two different situations is the job and what differs is only the scenery.
You locate it in the body first and name the quality: tight, hot, heavy, hollow, restless, cold. A located part can be sat beside and addressed, which makes an interview possible instead of a takeover, and attending to the body keeps answers coming from the layer the material is stored on rather than the layer you can already explain. That is Gendlin's felt sense, a method with modest support from small studies, alongside process research linking depth of experiencing to outcome across therapies. One caution. For some women, attention to the inside of the body increases distress. If that happens, open the eyes, look around you, feel the feet, and stop.
The fear question is a ladder: what are you afraid would happen if you stopped, then what would be so bad about that, then what would that mean about me. This is the downward arrow from cognitive therapy (Beck et al., 1979; Burns, 1980), which uses it to reach a core belief and then dispute it. This course uses the same ladder and stops at the top step, because a core belief held by a protective part is the material a head-on challenge hardens. You need to know what the belief is, in its own words, because you cannot meet what you have not found. Modules 3 to 5 work on it by a different route.
Then you ask the critic how old it thinks you are. A protective part can hold an assessment formed when it took the job and go on operating from it without registering that conditions have changed: that you now have your own money, that you can walk away, that you are forty-eight and not nine. If a part governing how you speak in meetings believes you are eleven, its behaviour becomes coherent immediately, and the gap between its number and the number on your passport is the size of the update that has not happened. This is clinical method across several traditions (IFS, Janina Fisher's work with self-alienated parts, van der Hart, Steele and Nijenhuis on structural dissociation, schema therapy's child modes) rather than a tested procedure. Treat the number as information about how the part is organised.
Naming which kind of critic was present converts an atmosphere into a describable presence with a known job. While you are blended you see through the part's eyes and its view feels like plain reality; once it is named there is a you doing the naming. Take the first answer. Labelling the emotion on a face reduces the amygdala response and raises activity in one prefrontal region, the right ventrolateral prefrontal cortex (Lieberman et al., 2007). That is suggestive, though the study names emotions rather than parts and the transfer should not be overclaimed.
Last, before any recognition is offered, the practice asks how you feel toward the part right now: warmth or curiosity, blankness, or something hot. This is the Self-energy check from IFS, and it decides which of two different events happens next. Words of recognition delivered from contempt do not soothe a protector. They tell it you want it gone and are being polite about it. The check stops you performing a warmth the part would detect instantly, and a hot or blank answer is information, because a second part has just made itself known. You are not asked to shift what you found, only to know it before you speak.
What we are actually doing. Recognition is offered in two separate parts, each in two wordings. The first is for the intent: what the part took on when it was young, with what it had. The second is for the weight: how long it has held the job, and that it held it alone. Each can be said as thanks ("thank you for trying to keep me safe") or as recognition without thanks ("I recognise what you were trying to do. I see your intention"), and the practice states plainly that neither wording is more correct. Then you check your chest and notice what happened, including nothing.
Then, deliberately, comes the resistance step. Two objections are named out loud and neither is argued with. That thanking it means it wins and the words it used were fair. And that softening toward it will make it worse, louder, or will collapse you, because the hardness is what has been holding you upright. Both are invited to be heard rather than resolved.
Then a floor, for the woman who found no warmth at all. Three versions, each complete: willingness instead of warmth ("I am willing to understand what you were doing"), two true statements held together and neither of them kind ("I would not have chosen you" and "you have not had a day off since I was a child"), or one breath with a hand on the chest and no words.
Why the practice is built around the resistance rather than the warmth. This is where the evidence is most direct, and it deserves stating plainly. Fear of self-compassion is a common, measurable pattern, and its association with mental health difficulty is significantly stronger in clinical samples than in non-clinical ones.
Gilbert, McEwan, Matos and Rivis (2011) developed and validated the Fears of Compassion scales, which measure fear of giving compassion, of receiving it from others, and of directing it toward yourself. Fear of self-compassion is associated with self-criticism, insecure attachment and depression. Kirby, Day and Sagar (2019), a meta-analysis in Clinical Psychology Review, pooled this literature and found that the more someone fears self-compassion, the more mental health difficulty they tend to report, with shame, self-criticism and depression among the strongest associations. Fearing warmth from other people tracks about as closely. Fearing giving it to others tracks much less. These are associations measured at a single point rather than predictions about what follows. Kelly, Carter, Zuroff and Borairi (2013), in Psychotherapy Research, followed patients through specialised eating-disorder treatment. Those who were both low in self-compassion and high in fear of it showed no meaningful improvement in shame or symptoms across the treatment. It is a preliminary study, and the finding is about that combination rather than about fear on its own. So the fear carries a clinical cost as well as a felt one: it predicts poorer response to treatment.
There is a physiological finding alongside the self-report. Rockliff, Gilbert, McEwan, Lightman and Glover (2008) gave participants a compassion-focused imagery exercise and measured heart-rate variability and cortisol. Some participants, tentatively those higher in self-criticism and insecure attachment, showed reduced heart-rate variability, a threat-like pattern, in response to imagining being cared for. Their bodies read incoming warmth as a threat rather than as comfort. It was a small pilot study and should be held as such, though the direction of the finding matches what clinicians see week after week.
Read together, this says something specific about how to build a practice like this one. For a meaningful proportion of self-critical women, an instruction to feel warmth toward yourself is asking for the exact response their system is defended against, and delivering it without preparation produces threat, shame at not being able to do it, and dropout. So the module names the objections before you can conclude something is wrong with you for having them, offers wordings that do not require gratitude, and provides a floor that requires no warmth at all. The resistance is treated as expected, informative, and workable, because that is what the evidence says it is.
The same finding explains why the workbook names the fear of letting go in words and then turns to the body. The fear that stops a woman releasing the critic is rarely held as a sentence. It sits as a bodily conviction, below language, that something catastrophic follows if the guard comes down. Attending to where it sits, at a slight distance so you stay inside what you can hold, works on the layer the fear lives on. The distance is the safety, which is why that pass is short and hedged with grounding.
On the form the recognition takes. The practice borrows the shape of thanks, and it is useful to be exact about where the warrant comes from. It comes from parts work, where turning toward a protector with recognition rather than opposition is the core move (Schwartz, 2021; Schwartz and Sweezy, 2020), and from compassion-focused and self-compassion work, where a warmer internal stance is associated with lower self-criticism and better wellbeing (Neff, 2003; MacBeth and Gumley, 2012; Gilbert, 2009).
It does not come from the gratitude literature, and that literature should not be cited for it. The gratitude research, including Emmons and McCullough's well-known work, studies gratitude directed toward other people, toward circumstances, or toward life in general. None of it studies gratitude directed toward a part of yourself. There is no evidence that findings about thanking people transfer to thanking an internal protector, and it would be sloppy to imply otherwise. Gratitude is the form being borrowed. Parts work and compassion-focused work are the evidence base. The practice only makes the second claim.
Used across: IFS (appreciating the protector for its work as a standard step before asking it to try anything different); CFT (building the soothing system rather than arguing with the threat one, with fear of compassion addressed directly as part of the work); mindful self-compassion (Neff, 2011; Germer and Neff, 2019), which also treats backdraft, the surge of difficult feeling that can follow self-kindness, as an expected event rather than a complication.
Why it works. Recognition allows a protector to ease, and naming the resistance first is what keeps recognition from becoming one more demand you fail.
The optional practice - its own audio and PDF, in the section above - has two steps, and they sit at very different distances from the evidence.
Step A - Letting it rest. Parts work and self-compassion, nothing more exotic. Turning toward a protector with recognition and warmth rather than argument is the core move of Internal Family Systems (Schwartz, 2021) and of mindful self-compassion (Neff, 2011; Germer and Neff, 2019), and the finding across this literature is a strong association between higher self-compassion and lower psychological distress (MacBeth and Gumley, 2012). The studies measure both at one time, so they show association rather than cause. Step A offers one moment of rest to a part that has been on duty since childhood. The claim to hold is "warmth does more than harshness, and this is a felt practice", not "this technique is proven". It is no substitute for therapy.
Step B - What came down the generations. This comes from Family Constellations, a group method developed by Bert Hellinger, and it falls outside what clinical psychology can firmly test. A systematic review by Konkolÿ Thege and colleagues (2021) in Family Process found only a small number of quantitative studies with few participants between them. Most reported improvement afterwards, and pooling a handful of them showed a moderate effect on general distress. The authors were clear that the evidence base is small and that the overall quality of the evidence is low, so those numbers do not establish that the method works. Rather over half the studies looked at harms, and several found minor or moderate negative effects in a small proportion of participants. More recent controlled work is still limited and mixed. Much of Hellinger's own framework, above all the claim that the exercise transmits real information about real ancestors, sits beyond what any study can test.
That is why the module keeps Step B strictly imaginal and self-authored. It does not import Hellinger's "orders of love", his reconciliation doctrine, or any claim that the reader is receiving information about actual people. It uses one element only, the felt image of a weight travelling down through the generations and being set down, because for some women that image does something a plain explanation does not. It is offered with a boundary block at the top, grounding before and after, and a therapist signpost at the close.
Neither step builds the inner ally. Resting a protector and setting down an inherited pattern are single, present-moment moves. The voice that keeps meeting the critic is built in Module 3, deliberately and over time.
You have asked the critic what it has been trying to do, and you have seen the job underneath the harshness. For some women this module ends the fight. For others it opens anger, or grief for the child who learned to attack herself first, and none of that needs to be resolved this week. Next module we begin building the inner ally, the voice that can hold what the critic cannot. For this week, the function work is enough.
Module 3 · Meeting Your Inner Ally →