Argyris Double-Loop Learning

Document espoused values and policies and compare against observed behaviours and decisions to surface inconsistencies.

• values statement
• behavioural audit reports
• 360 feedback
• decision case studies

• no behavioural audit
• values aspirational only
• leaders exempt from review

Identify and document Model I defensive routines (unilateral control, win not lose, suppress feelings, rationality) that block learning.

• defensive-routine inventory
• facilitator notes
• action science case write-ups

• defensiveness not named
• executives protected
• no facilitator training

Adopt Model II behaviours (valid information, free informed choice, internal commitment) in decision and review forums.

• meeting protocols
• Model II training records
• facilitator certifications

• Model II training one-off
• no facilitator pool
• reverts under stress

Ensure lessons learned from incidents, audits, and reviews modify governing variables, not just operational responses.

• policy change log linked to lessons
• governing variable register
• audit closeout reports

• lessons not actioned
• policy never updated
• loops never close

Measure organizational climate for psychological safety, voice, and willingness to challenge governing assumptions.

• climate survey
• trend analysis
• focus group notes

• survey unread
• scores ignored
• retaliation feared

Balancing Advocacy with Inquiry - stating views while genuinely inviting challenge and alternative perspectives. Control from Argyris Double-Loop Learning framework, domain: Defensive Routines and Learning.

• After action review template and records
• Learning forum minutes with surfaced assumptions
• Process owner attestation
• Tooling configuration export
• Meeting minutes referencing the control
• Training material referencing the control

• Learning stays at single loop without challenging assumptions
• Defensive routines block honest feedback
• Evidence is point in time rather than ongoing
• Control owner unclear or vacant
• No metric tracks control effectiveness

Organizational Defensive Routines - actions preventing embarrassment or threat that also prevent learning. Control from Argyris Double-Loop Learning framework, domain: Defensive Routines and Learning.

• After action review template and records
• Learning forum minutes with surfaced assumptions
• Process owner attestation
• Tooling configuration export
• Meeting minutes referencing the control
• Training material referencing the control

• Learning stays at single loop without challenging assumptions
• Defensive routines block honest feedback
• Evidence is point in time rather than ongoing
• Control owner unclear or vacant
• No metric tracks control effectiveness

Ladder of Inference - awareness of how people move from data to conclusions, often skipping steps. Control from Argyris Double-Loop Learning framework, domain: Defensive Routines and Learning.

• Training completion records by role
• Phishing simulation results
• After action review template and records
• Learning forum minutes with surfaced assumptions
• Process owner attestation
• Tooling configuration export

• Role based training not delivered to high risk teams
• Training metrics not reported to leadership
• Learning stays at single loop without challenging assumptions
• Defensive routines block honest feedback
• Evidence is point in time rather than ongoing

Left-Hand Column Exercise - surfacing what people think but do not say to reveal gaps between espoused and actual. Control from Argyris Double-Loop Learning framework, domain: Defensive Routines and Learning.

• OT asset inventory and network diagram
• ICS specific incident playbook
• After action review template and records
• Learning forum minutes with surfaced assumptions
• Process owner attestation
• Tooling configuration export

• IT and OT response teams not aligned
• ICS forensics tooling not in place
• Learning stays at single loop without challenging assumptions
• Defensive routines block honest feedback
• Evidence is point in time rather than ongoing

Productive Inquiry and Conversation - creating conditions for honest dialogue without defensiveness. Control from Argyris Double-Loop Learning framework, domain: Defensive Routines and Learning.

• After action review template and records
• Learning forum minutes with surfaced assumptions
• Process owner attestation
• Tooling configuration export
• Meeting minutes referencing the control
• Training material referencing the control

• Learning stays at single loop without challenging assumptions
• Defensive routines block honest feedback
• Evidence is point in time rather than ongoing
• Control owner unclear or vacant
• No metric tracks control effectiveness

Board agenda includes periodic review of whether governing values, risk appetite, and strategic assumptions still hold.

• board minutes
• strategy refresh papers
• assumption review pack

• assumptions never revisited
• strategy reaffirmed by default

Use action science cycle to intervene, test interventions, and revise theories of action based on observed behaviour change.

• intervention plan
• baseline observation
• post-intervention measurement
• theory revision notes

• no baseline
• intervention not measured
• theory unchanged

Detect cases where senior practitioners produce outcomes opposite to intended through highly skilled defensive behaviour.

• leadership 360 reports
• case studies
• executive coaching logs

• senior leaders exempt
• feedback sanitised
• no remediation path

Periodically audit executive reasoning patterns for consistency with learning principles and willingness to surface assumptions.

• recorded decision sessions
• reasoning audit report
• remediation plan

• leaders refuse audit
• audit cosmetic
• no remediation

Double-Loop Learning - questioning and modifying underlying assumptions, values, and policies. Control from Argyris Double-Loop Learning framework, domain: Learning Loops.

• Signed and dated policy set with version history
• Annual review and approval records
• After action review template and records
• Learning forum minutes with surfaced assumptions
• Process owner attestation
• Tooling configuration export

• Policies past their review date
• No evidence policies were communicated to staff
• Learning stays at single loop without challenging assumptions
• Defensive routines block honest feedback
• Evidence is point in time rather than ongoing

Single-Loop Learning - correcting errors within existing governing variables without questioning assumptions. Control from Argyris Double-Loop Learning framework, domain: Learning Loops.

• After action review template and records
• Learning forum minutes with surfaced assumptions
• Process owner attestation
• Tooling configuration export
• Meeting minutes referencing the control
• Training material referencing the control

• Learning stays at single loop without challenging assumptions
• Defensive routines block honest feedback
• Evidence is point in time rather than ongoing
• Control owner unclear or vacant
• No metric tracks control effectiveness

Triple-Loop Learning - learning about how we learn, transforming the learning system itself. Control from Argyris Double-Loop Learning framework, domain: Learning Loops.

• After action review template and records
• Learning forum minutes with surfaced assumptions
• Process owner attestation
• Tooling configuration export
• Meeting minutes referencing the control
• Training material referencing the control

• Learning stays at single loop without challenging assumptions
• Defensive routines block honest feedback
• Evidence is point in time rather than ongoing
• Control owner unclear or vacant
• No metric tracks control effectiveness

Establish capacity to detect and correct errors against existing norms, policies, and operating assumptions without questioning the underlying frame.

• post-incident review records
• corrective action tracker
• KPI dashboards
• root cause analyses

• fixes treat symptoms not causes
• no closure tracking
• metrics not tied to learning

Define thresholds and signals that escalate a problem from single-loop correction to examination of governing values, assumptions, and policies.

• escalation policy
• repeat-incident reports
• governance review minutes
• assumption-challenge logs

• no escalation trigger defined
• repeat incidents not aggregated
• assumptions never written down

Train leaders and teams to surface data selection, meaning, assumptions, and conclusions before acting on inferred beliefs.

• ladder of inference training materials
• decision logs with assumption tests
• peer challenge records

• assumptions implicit
• challenge avoided
• no decision log

Use left-hand column technique to expose unspoken thoughts and feelings that shape conversations and block learning.

• case write-ups
• coaching session notes
• redacted dialogue samples

• psychological safety low
• exercise treated as gimmick
• no follow-up

Require claims to be supported by illustrative data, reasoning made explicit, and inferences testable by others.

• decision paper template
• peer review criteria
• board pack standards

• assertions unsupported
• reasoning compressed
• no peer review

Map cross-functional defensive routines (mixed messages, undiscussability, cover-up) that protect the system from learning.

• defensive routine map
• undiscussable issue register
• action plan

• undiscussables remain undiscussed
• map filed not used

Espoused Theory - the theory of action people claim to follow, their stated values and beliefs. Control from Argyris Double-Loop Learning framework, domain: Theories of Action.

• After action review template and records
• Learning forum minutes with surfaced assumptions
• Process owner attestation
• Tooling configuration export
• Meeting minutes referencing the control
• Training material referencing the control

• Learning stays at single loop without challenging assumptions
• Defensive routines block honest feedback
• Evidence is point in time rather than ongoing
• Control owner unclear or vacant
• No metric tracks control effectiveness

Model I Behavior - unilateral control, win-lose framing, suppression of negative feelings, defensiveness. Control from Argyris Double-Loop Learning framework, domain: Theories of Action.

• AI system inventory with risk classification
• Model card and evaluation report
• After action review template and records
• Learning forum minutes with surfaced assumptions
• Process owner attestation
• Tooling configuration export

• Model performance drift not monitored
• Human oversight thresholds not defined
• Learning stays at single loop without challenging assumptions
• Defensive routines block honest feedback
• Evidence is point in time rather than ongoing

Model II Behavior - valid information, free and informed choice, internal commitment, bilateral control. Control from Argyris Double-Loop Learning framework, domain: Theories of Action.

• AI system inventory with risk classification
• Model card and evaluation report
• After action review template and records
• Learning forum minutes with surfaced assumptions
• Process owner attestation
• Tooling configuration export

• Model performance drift not monitored
• Human oversight thresholds not defined
• Learning stays at single loop without challenging assumptions
• Defensive routines block honest feedback
• Evidence is point in time rather than ongoing

Theory-in-Use - the theory of action actually governing behavior, often unconscious and different from espoused. Control from Argyris Double-Loop Learning framework, domain: Theories of Action.

• After action review template and records
• Learning forum minutes with surfaced assumptions
• Process owner attestation
• Tooling configuration export
• Meeting minutes referencing the control
• Training material referencing the control

• Learning stays at single loop without challenging assumptions
• Defensive routines block honest feedback
• Evidence is point in time rather than ongoing
• Control owner unclear or vacant
• No metric tracks control effectiveness