Human and Machine: A Journey - Part 14
Reflection and Learning Systems
Throughout our journey exploring welfare systems, we've examined knowledge flows between research and practice, professional judgment within systematic frameworks, and the transformative potential of equality and youth engagement. Now we turn to perhaps the most fundamental element that sustains all these dimensions: reflection and learning—the processes through which both individuals and organizations develop.
My own journey with reflection on learning systems began unexpectedly when I was just 15 years old. While many of my classmates found themselves in typical internships at shops, care facilities, or offices, I had the unusual opportunity to spend a week in a molecular biology laboratory. The professor running the lab was a family friend, and what seemed at first like an intimidating environment quickly became a fascinating window into a different world of learning.
I remember vividly the delicate work of transferring tiny C. Elegans worms between bacterial plates. Using a small metal pick under a microscope, I had to develop just the right touch—firm enough to lift the nearly microscopic creatures, but gentle enough not to crush them. Each attempt brought immediate feedback: either a successfully transferred worm or a small smudge that represented failure. No manual or classroom lecture could have taught this tacit knowledge that resided in the fingertips.
More profound than the laboratory skills, however, was a conversation with the professor that has stayed with me for decades. Looking up from his microscope, he observed how strange our educational system really is:
We force everyone through the same formalized path, year after year in classrooms. Yet history's most groundbreaking thinkers—Einstein, Darwin, Curie—often began their scientific journeys as children through apprenticeships and direct mentoring. They learned by doing, observing, and being guided by masters in their field.
This insight illuminates our ongoing exploration of human and machine approaches. Traditional education systems, with their standardized curricula and assessment methods, represent the machine approach to learning—systematic, uniform, and optimized for efficiency. The apprenticeship model represents the human approach—personalized, relationship-based and centered on tacit knowledge transmission through observation and practice.
The Reflecting Team: A Bridge Between Human and Machine
This tension between standardized and experiential learning manifests powerfully in welfare services. How do professionals develop the sophisticated judgment needed for complex human work within increasingly standardized systems? How do organizations learn from experience without defaulting to purely mechanical metrics?
Tom Andersen's groundbreaking work on the Reflecting Team approach offers a powerful framework for navigating this challenge. Developed in family therapy contexts during the 1980s, Andersen's approach creates a structured process that simultaneously honors both systematic rigor and human wisdom.
In a typical Reflecting Team session:
Practitioners first engage directly with clients (participation)
They then step back to observe and reflect with colleagues (observation)
This reflection happens openly, with clients present
Finally, clients reflect on the team's reflections
This elegant process creates what we might call a reflective loop - moving systematically between participation and observation, between direct experience and thoughtful analysis of that experience. It provides structure without rigidity, creating what Andersen called appropriate difference - enough change to stimulate new insights but not so much as to overwhelm.
When I first encountered Andersen's approach early in my career, it transformed my understanding of professional learning. Here was a method that didn't position theory and practice as opponents but as dance partners, each enriching the other through structured dialogue. The approach perfectly embodied the integration of human and machine elements we've explored throughout this series—combining systematic process with human connection.
Beyond Methods: Creating Learning Systems
While specific methods like the Reflecting Team approach provide valuable tools, creating genuine learning organizations requires broader systemic changes. Drawing from my experience developing welfare organizations, several key elements emerge as crucial for establishing what we might call reflective infrastructure:
Protected Reflection Spaces
First, organizations must create both physical and temporal spaces where reflection can occur without immediate practical demands. These protected spaces include:
Formal structures - like regular supervision sessions, case reflection meetings, and learning circles where practitioners can systematically analyze complex cases.
Informal opportunities - like hallway conversations, professional friendships, and communities of practice where tacit knowledge and professional wisdom can emerge organically.
Making Tacit Knowledge Visible
Second, effective learning systems develop processes for making tacit knowledge visible and shareable. The fingertip knowledge I developed working with C. Elegans worms represents precisely the kind of tacit understanding that professional welfare work demands—knowledge that exists in practice but resists simple codification.
Several approaches help make this tacit knowledge more accessible:
Narrative documentation - that captures not just what happened but why practitioners made specific choices. When social workers document not just assessment outcomes but their reasoning process, they make professional judgment more visible and therefore more available for collective learning.
Structured dialogues - where experienced practitioners externalize their thinking, making their internal decision-making processes available to others. Methods like the reflecting team in team group meetings discussing case strategies.
Metaphor and visualization - techniques that translate hard-to-articulate knowledge into more accessible forms. When practitioners create visual representations of their understanding or relate their understanding to images, they often surface insights that remained implicit in purely verbal exchanges.
These approaches don't mechanize tacit knowledge but rather create bridges between tacit and explicit understanding, allowing organizations to learn from their collective wisdom rather than just individual experience.
Developmental Evaluation
Third, learning systems require evaluation approaches that support development rather than just accountability. Traditional performance metrics often represent the machine approach at its most limiting—reducing complex professional judgment to simplified indicators that can inadvertently distort practice.
Developmental evaluation offers an alternative that maintains necessary rigor while supporting learning and adaptation. This approach:
Positions evaluation as an ongoing process rather than an occasional event
Focuses on generating insights for improvement rather than just proving compliance
Engages practitioners as active participants rather than passive subjects
Adapts evaluation questions and methods as understanding evolves
Maintains methodological flexibility while ensuring systematic inquiry
The Hourglass Reimagined: Learning Across Levels
When these elements come together, they fundamentally transform the hourglass model of knowledge flow we've explored throughout this series. Traditional models position research knowledge at the top, flowing down to practice, with limited channels for practice knowledge to influence research and policy.
Learning systems create multi-directional knowledge flows, where insights from practice need to systematically inform both policy and research. We need research with both quantitative methods and qualitative practice-based evidence that reveals how interventions actually work in complex real-world contexts.
Imagine a child welfare organization that implements a new evidence-based assessment framework. In a traditional model, practitioners simply implement the framework as designed. In a learning system:
Practitioners document both successes and challenges in implementation
Regular reflection sessions identify patterns across different contexts
Emerging insights shape ongoing adaptation of the framework
Modifications are systematically tested rather than implemented ad hoc
Learning from local adaptation feeds back to researchers and policymakers
This multi-directional flow creates what we might call a learning ecosystem rather than a linear knowledge pipeline. It maintains the necessary structure of systematic inquiry while creating space for the human elements of professional wisdom to shape development.
Digital Learning Systems: New Possibilities
The digital transformation of welfare services brings both challenges and opportunities for reflective practice. Digital tools can either constrain or enhance professional reflection, depending on how they're designed and implemented.
Many current documentation systems prioritize structured data capture over reflective documentation, asking practitioners to categorize observations into predetermined fields rather than capture the rich complexity of their professional reasoning. These systems, designed primarily for administrative accountability, can inadvertently reduce rather than enhance learning.
Yet digital tools could potentially transform reflective practice. Imagine systems that:
Use natural language processing to analyze narrative documentation for patterns and insights
Create visual representations of complex case situations that reveal otherwise hidden connections
Support asynchronous reflecting team processes across distributed organizations
Provide real-time access to relevant research and practice wisdom during complex decisions
Use machine learning to identify successful adaptations that might benefit other contexts
The Meta-Challenge: Learning About Learning
Perhaps the most profound challenge facing welfare organizations is what we might call the meta-challenge - learning how to learn more effectively. This requires developing what Chris Argyris termed double-loop learning - not just improving existing approaches but questioning the underlying assumptions that shape those approaches.
In welfare services, double-loop learning means moving beyond questions like How can we implement this assessment framework more effectively? (single-loop) to questioning the underlying goals and assumptions themselves: Why are we using this framework in the first place? Are the goals and assumptions embedded in this framework truly the most appropriate for helping families? Should we reevaluate our entire assessment philosophy? This approach recognizes that the way we define and solve problems may itself be the source of the problem.
This meta-level reflection connects directly back to the professor's insight about education systems that I encountered in that laboratory years ago. Traditional learning often focuses on acquiring established knowledge and skills - single-loop learning within existing frameworks. Transformative learning requires questioning those very frameworks - recognizing when the traditional paths might be limiting rather than enhancing our understanding.
Several approaches support this meta-level reflection:
Critical reflexivity - that explicitly examines how our social positions and cultural assumptions shape both our practice and our understanding of that practice. This goes beyond technical reflection on what works to examine the values and power dynamics that define "working" in the first place.
Cross-disciplinary dialogue - that brings different knowledge traditions into conversation, revealing assumptions that remain invisible within any single tradition. When child welfare practitioners engage with family therapists, community organizers, psychologists and doctors, they encounter fundamentally different ways of understanding family support that can challenge taken-for-granted approaches.
Historical analysis - that examines how current practices emerged over time rather than treating them as natural or inevitable. Understanding that many current welfare approaches developed in specific historical contexts helps practitioners recognize possibilities for fundamentally different approaches rather than just incremental improvements.
These approaches don't reject systematic knowledge but rather enrich it by revealing its contextual nature and opening spaces for alternative understandings that might better serve complex human needs.
From Machine Learning to Human Learning
As artificial intelligence transforms welfare services, the question of learning becomes even more crucial. While much attention focuses on how machines learn from data, equally important is how humans learn alongside increasingly sophisticated systems.
This raises fundamental questions about human-machine learning partnerships:
How do professionals develop judgment when algorithms handle increasing portions of decision processes?
What new forms of tacit knowledge emerge in AI-augmented welfare work?
How do organizations learn from the interaction between human judgment and algorithmic insights?
What reflection processes help practitioners understand the strengths and limitations of AI systems?
These questions don't have simple answers, but they point toward new forms of reflective practice that explicitly address human-machine interaction rather than treating technology as simply a tool. Just as the delicate work with C. Elegans required developing tacit knowledge through direct experience, working effectively with AI systems requires similar development of judgment that goes beyond technical understanding.
Building Learning Organizations in Practice
My own experience developing reflective infrastructure in welfare organizations has taught me that success requires harmonizing structure and flexibility. Let me share how we created a unifying approach that transformed our organization:
When working with our social services team in a rural municipality, we recognized that traditional implementation approaches often created tension between standardization and professional autonomy. Rather than pursuing stricter compliance or accepting complete individualization, we developed what we called a framework for consensus (in the methaphorical model of a house - check out this interactive model) – a holistic organizational model that established clear boundaries while honoring individual variation.
This framework wasn't just another assessment tool – it became the integrating structure for our entire organization. We deliberately designed it to be comprehensive yet adaptable, touching every aspect of our operations:
Quality management became linked to the framework's core values and principles, providing consistent evaluation criteria
Reporting to the municipal board utilized the framework's structure, creating alignment between practice and governance
Weekly communications within the organization followed the framework's categories, reinforcing its relevance to daily work
Professional development and salary criteria were integrated with the framework, connecting individual growth to organizational priorities
What made this approach powerful was its ability to create unity without uniformity. The framework established clear parameters and shared values, but encouraged variation within those boundaries. Staff weren't expected to practice identically – they were invited to contribute their unique perspectives while maintaining alignment with core principles.
Particularly significant was our development of this approach in a rural context, where resources are limited and versatility is essential. We found that our framework created a balance that suited our environment – structured enough to ensure consistency but flexible enough to adapt to the unique challenges of rural social work.
The results transformed our organizational culture: reduced conflict around competing values, increased staff engagement in development work, and a shared language that facilitated both learning and accountability. Most importantly, practitioners developed what I call principled adaptation – the ability to maintain fidelity to core values while thoughtfully customizing approaches to meet diverse family needs.
I believe this approach offers a transferable model that other organizations can adapt to their contexts. The key is finding the right level of specificity – detailed enough to provide real guidance but open enough to accommodate professional wisdom and local conditions.
Looking Forward: From Learning Systems to Learning Societies
As we move toward Part 15's exploration of connecting research and practice, consider how our reflections on learning might extend beyond individual organizations to welfare systems as a whole. What would it mean to create not just learning organizations but learning societies where multiple forms of knowledge - professional, research-based, experiential, and cultural - continuously interact to create more effective and humane welfare approaches?
This question brings us full circle to my experience in that laboratory decades ago. The professor's insight about education systems points toward deeper questions about how societies organize learning and knowledge development. Traditional education systems with their formalized, sequential structures represent one approach. The apprenticeship model that guided Darwin, Einstein, and countless other innovative thinkers represents another.
The challenge for welfare systems isn't choosing between these approaches but finding ways to integrate their strengths. We need both the systematic rigor of formal knowledge and the tacit wisdom that comes from direct experience and mentoring. We need both structured processes that ensure consistency and flexible approaches that adapt to complexity.
As we navigate increasingly complex social challenges and rapidly evolving technologies, this integration becomes not just desirable but essential. The welfare systems that thrive will be those that create sustained dialogue between different knowledge forms, that build reflective infrastructure at multiple levels, and that value learning as core to their mission rather than an occasional luxury.
Before we explore connections between research and practice in Part 15, consider:
What helps you move between participation and observation in your professional practice?
How does your organization capture and share tacit knowledge?
What structures support reflection amid increasing demands for efficiency?
How might digital transformation create new possibilities for reflective practice?
This is part 14 in our ongoing series exploring the intersection of human judgment and systematic knowledge in modern welfare systems. Join the conversation by sharing your thoughts and experiences in the comments below.