Welfare Footprint Institute

Defining Affective Experiences

editar — Fri, 11 Apr 2025 18:26:48 +0000

How to Define Affective Experiences for Analysis: Striking the Right Level of Detail

Cynthia Schuck-Paim & Wladimir Alonso

April 11, 2025

The Welfare Footprint Framework (WFF) is designed to quantify the welfare impact of each and every affective experience that the target animals endure over a timeframe of choice, as measured in terms of its intensity (how pleasant, painful, frustrating, or distressing it is) and duration (how long it lasts). The goal is to understand and compare the welfare of animals (i.e., the cumulative time spent in affective states of different intensities) across conditions and systems in a way that is both scientifically sound and useful. To do this, the WFF relies on a set of building blocks (note: we use initial capital letters for terms with operational definitions, described in this glossary):

Circumstances: the environmental and management conditions that animals encounter
Biological Outcomes: the immediate consequences resulting from an organism’s interaction with Circumstances. They can be positive or negative, physical or psychological. Not all Biological Outcomes are affectively relevant: for example, immune suppression is not ‘felt’, but is an intermediate step that may lead to an affectively relevant Biological Outcome (a painful disease).
Pathways: each Biological Outcomes may unfold differently in different animals, leading to a different Affective Experience. These different progressions are referred to as Pathways. Common Pathways include: Acute vs. chronic, Fatal vs. non-fatal, Mild vs. severe.
Affective Experiences: These are the felt states that the WFF aims to quantify—such as pain, fear, frustration, comfort, pleasure, or satiety.

In conducting any welfare impact analysis, one must first identify the multiple Affective Experiences that animals endure over their lives in the different scenarios analysed. The purpose of this text is to help users define which Affective Experiences should be analyzed and at what level of specificity. This involves finding a balance between biological plausibility, data availability, and the usefulness of the results.

How to Decide What Level of Detail Is Appropriate

One of the most critical steps in applying the Welfare Footprint Framework is deciding how Affective Experiences should be defined for analysis. This step determines the resolution of the welfare quantification—too broad, and important nuances may be lost; too detailed, and the analysis may become unmanageable or unsupported by data. The challenge lies in avoiding both oversimplification and unnecessary complexity.

Not enough detail

It is important to recognize that the appropriate level of detail will depend on the objective of the analysis. In some cases, the goal may require rapidly estimating the overall affective state of an animal during a given phase of life, such as the rearing or production phase. In such cases, it may be valid to use broad categories of experience—such as general discomfort or well-being—without distinguishing between every individual disease, deprivation, or positive event the animal may encounter. However, this type of broad estimation cannot be directly supported by empirical evidence on intensity and duration. That is because the scientific literature and field data provide validated indicators of intensity and duration only for specific conditions, such as fractures, respiratory infections, social bonding, or restricted movement.

To ensure that the analysis remains evidence-based, it is often necessary to define affective experiences at a more granular level. This means anchoring them in specific biological outcomes and pathways that are known to result in distinct experiences. Pain from a fractured limb, pleasure from affiliative interactions, or fear from predator exposure are all examples of experiences with documented affective relevance and measurable properties. These experiences are not only more scientifically tractable—they also tend to mark events that are especially meaningful in the life of the animal. They help characterize key welfare risks and opportunities.

Too much detail

At the other end of the spectrum, one may be inclined to define affective experiences in an overly detailed manner. This often happens when attempting to distinguish between every minor variation in symptom presentation or disease etiology, believing that more granularity necessarily equates greater accuracy. However, excessive granularity can make the analysis cumbersome and can exceed the resolution supported by available data. For example, differentiating between every possible cause of lameness, or treating slight differences in respiratory pathogen strains as separate experiences, may not yield meaningful distinctions in what the animal feels.

A similar issue arises when users define affective experiences based on symptoms (e.g., fever, coughing). This approach may appear detailed or clinically grounded, but it introduces unnecessary complexity and limits the usefulness of the analysis. Symptoms can indeed be part of an affective experience—such as coughing that results from a painful airway irritation —but they are often not reliable units of analysis on their own. Symptoms reflect observable aspects of a disease process, and their presence may suggest a certain intensity or progression.

Moreover, symptoms (or indicators – see this figure) do not provide a dependable basis for quantification. There is often no validated data on the duration and prevalence of individual symptoms in a population. For example, we might know how many animals suffer from tracheitis, and for how long, but we typically lack systematic data on how many animals cough, how frequently, and for how long in a way that can be used for quantification. Because symptoms are part of the unfolding of a disease over time, they are best viewed as indicators or inputs—useful in clinical judgment or diagnostic pathways. Additionally, the same symptom can appear across a range of very different conditions, and its affective relevance depends entirely on context. Coughing could reflect a temporary irritation or a necrotizing infection. Treating symptoms as affective experiences overlooks this variability and masks the differences in what the animal may actually feel.

Classifications based on Aetiology

Another common tendency is to define Affective Experiences in terms of the specific aetiology of a condition—for example, as “respiratory disease caused by respiratory syncytial virus (RSV),” or “respiratory disease caused by influenza virus,” or “respiratory disease caused by metapneumovirus.” While this level of detail might reflect a clinical interest in diagnosis, it is not helpful for affective experience analysis. First, population-level data on the prevalence of these specific viral causes is rarely available, as diagnostic testing is typically not performed routinely, especially in large-scale production systems. Second, these aetiologies often result in highly similar symptoms, and the affective experience they produce (such as irritation, air hunger, or fatigue) is often indistinguishable. In human health, a common practice is to refer to these overlapping presentations as “Influenza-like illness” (ILI), recognizing that the syndrome is more relevant to the patient’s experience than the precise pathogen. The same logic applies here: rather than defining separate affective experiences by each pathogen, it is better to define them at the level of the syndrome or biological outcome they share. This allows for more realistic, evidence-based estimates of intensity and duration, and avoids artificial fragmentation of experience categories.

Some Useful Rules

Below we provide a set of principles to help users define affective experiences at a level of detail that is analytically sound, biologically meaningful, and operationally feasible for population-level welfare assessment

1. Group when experiences are similar in intensity and duration

Group biological outcomes into one affective experience only when they are likely to feel the same to the animal.

Example: Several mild respiratory infections that all cause brief mucosal discomfort might be grouped.

2. Keep separate when affective experiences are qualitatively distinct

Even if two biological outcomes affect the same body system, they should be treated as separate affective experiences if the animal is likely to feel them differently.

Example: Pain from a clean, well-aligned bone fracture should not be grouped with pain from a bone fracture with non-union that never heals and leads to chronic inflammation, even though both are classified as ‘bone fractures.’ The intensity of the pain caused by each type of fracture differ significantly. Defining them separately allows for more informative assessments, especially when comparing different housing systems, genetics, or management practices that affect the likelihood of one type over another.

3. Base decisions on the level of detail supported by data

Sometimes, the theoretically ideal distinctions cannot be made because population-level data is not available. In such cases, broader groupings may be necessary.

Example: Lameness in broilers can result from many causes. However, even when the underlying aetiology is the same, different birds may experience it in very different ways depending on the severity, location, and progression of the condition. Two animals with the same musculoskeletal issue might have completely different levels of discomfort, mobility restriction, or pain duration. Because welfare assessments often rely on gait scores, which integrate these aspects into an observable indicator, defining affective experiences based on gait score is both more feasible and more informative than trying to isolate every underlying etiology.

4. Make distinctions that are informative for welfare analysis

A key reason to define experiences separately is that it makes the quantification informative. If the goal is to compare the welfare consequences of two or more procedures, such as different vaccination methods, or environmental enrichments, the underlying affective experiences must be analyzed separately. Even if the tissue affected is the same or the outcome is similar in general terms, differences in how the condition is produced can lead to distinct experiences for the animal.

Example: Pain from thermal dehorning with a hot iron should not be grouped with pain from chemical dehorning, even though both are procedures applied to the same area of the animal’s body. The mechanism of tissue damage, progression of healing, and likelihood of chronic pain can differ significantly between these techniques. Defining these experiences separately is also essential for evaluating and comparing their welfare consequences .

5. Avoid grouping based on symptoms alone

Symptoms such as coughing, lethargy, or poor feather condition may result from a wide range of biological outcomes and should not be used as the basis for defining affective experiences. A clear diagnosis, including an understanding of cause, location, and mechanism, is needed.

Less Appropriate: Grouping all cases involving coughing as “Pain from coughing.”

More Appropriate: Defining pain based on the underlying cause, such as “Pain from upper respiratory tract infection.”

Confounding Factors in Welfare Comparisons

editar — Fri, 14 Mar 2025 16:13:38 +0000

Confounding Factors in Welfare Comparisons of Animal Production Systems

Wladimir J Alonso, Cynthia Schuck-Paim

March 14, 2025

Controlled experiments—where variables are deliberately manipulated to establish cause-and-effect relationships—are the gold standard for drawing reliable conclusions in science. However, controlled experiments are not always feasible, ethical, or practical, and one must instead rely on retrospective analyses of existing datasets to uncover meaningful patterns.

This is particularly common in social sciences, psychology, public health, epidemiology, and economics, where large observational datasets are often used to infer causality. For example, in public health, studies on the long-term effects of smoking or the impact of pollution on respiratory diseases have relied on existing medical records and environmental data rather than randomized trials, which would be unethical.

A classic example of how careful analysis of observational data can overturn conventional wisdom is the work of Daniel Kahneman, whose research on cognitive biases and decision-making (Thinking, Fast and Slow) demonstrated that humans do not always behave rationally, as previously assumed in classical economic models. His work, which earned him a Nobel Prize in Economics, has often relied on retrospective analyses of decision patterns rather than controlled trials.

Similarly, Hans Rosling, in Factfulness, debunked widespread myths about global poverty and development trends by analyzing historical datasets. His work demonstrated that when adjusted for factors like GDP growth, access to healthcare, and education levels, many claims about worsening global inequality were not supported by the actual data.

In economics, Thomas Sowell has extensively applied this principle in his analysis of income disparities, educational outcomes, and social mobility. His work demonstrates how controlling for factors like education level, work experience, and family structure can reveal patterns that differ from popular assumptions, emphasizing the need for data-driven, rather than narrative-driven, conclusions.

Transition to Animal Welfare Research

The same principles apply to animal welfare research, where controlled experiments are often unfeasible or unethical, and researchers must rely on industry data, automated monitoring systems, and observational studies to assess welfare outcomes. As the use of precision livestock farming and automated welfare monitoring technologies (such as computer vision tracking, behavioral sensors, and real-time physiological monitoring) expands, retrospective analyses of these large datasets will become even more critical in drawing valid conclusions about animal welfare in different production systems.

Without careful adjustments, comparisons may incorrectly attribute welfare outcomes to production systems rather than to confounding variables such as differences in management decisions, economic investment, types of animals or life-stage differences.

Here we describe some common confounding factors in welfare comparisons, and possible analytical approaches to address confounding.

Resource and Management Investment Disparities

If economic and human resources are allocated differently among production systems, welfare comparisons can be skewed. Feedlots, for example, typically invest more in veterinary care per animal than pasture-based systems. This can lead to lower mortality rates in feedlots, not because the system is inherently better for welfare, but because intensive monitoring and interventions can reduce the likelihood of acute health crises. However, these interventions may also mask underlying welfare issues of similar or worse severity in feedlots than in pasture-based systems.

ANIMAL SELECTION BIASES

Selection biases occur when different production systems use different genetic lines or selectively place healthier animals in specific systems. In the beef cattle industry, for example, grain-fed feedlot cattle are often selected from breeds with greater feed efficiency, or that, depending on market prices and the animals’ general health, are deemed worthy of the investment on a feedlot at the finishing phase. A comparison of welfare outcomes between these groups must therefore account for the pre-existing and underlying differences in the health and resilience between the two groups of animals.

Early Life Experience and System Adaptation

A critical confounding factor often overlooked is the animal’s developmental history and adaptation to specific environments. Production animals develop physiological and behavioral adaptations to their early rearing environments that significantly impact their ability to thrive when transferred to different systems later in life. Valid welfare comparisons must account for the congruence between rearing and production environments. A system that appears superior when populated with appropriately reared animals may perform catastrophically when housing animals with mismatched developmental histories. This understanding fundamentally challenges simplistic “system A versus system B” welfare comparisons. The welfare implications of any production system cannot be evaluated in isolation from the animal’s developmental trajectory. Meaningful comparisons must either match rearing environments to production systems or explicitly account for adaptation capacity as a welfare dimension in its own right.

Incomplete Welfare Outcome Coverage

A pervasive confounder in system comparisons is selective measurement of biological outcomes. Studies often focus on easily quantifiable biological outcomes like mortality, injuries, or production parameters while neglecting equally important biological outcomes such as behavioral and social deprivations or long-term health impacts. This selective inventory creates fundamental distortions in system evaluations. For example, early weaning practices in pig production may appear advantageous when measuring immediate post-weaning mortality and growth rates, while missing significant biological outcomes such as compromised gut development and impaired immune function. Valid welfare comparisons require clarity on the spectrum of biological outcomes considered across all relevant welfare domains in all systems considered.

Environmental and Geographic Variability

Welfare comparisons across production systems are frequently confounded by unaccounted environmental and geographic factors that operate independently of the system design itself. Production systems situated in different regions experience fundamentally different environmental challenges that significantly impact animal welfare outcomes. Climate variations, for instance, can affect the welfare implications of outdoor systems—the same pasture-based approach might provide excellent welfare in moderate climates but create severe thermal stress in regions with extreme temperatures. Valid welfare comparisons must therefore either control for environmental variables by comparing systems within similar geographic regions or explicitly incorporate environmental challenges as contextual factors in the analysis.

TEMPORAL CONSIDERATIONS

Welfare comparisons across systems that raise animals with different lifespans present another challenge. Short-lived production animals may experience fewer cumulative welfare issues simply because they do not live long enough for severe outcomes to develop, or because their longer lifespan leads to greater time spent in negative states. This can create a false perception that high-turnover systems lead to better health than systems where animals live longer. One solution is to compare the quality of life of animals on an average day. Where life changes over the production cycle, average day life quality can be assessed at different points in time. The average day approach normalizes welfare assessments across a standard time unit, ensuring that the overall quality of life is assessed rather than just its duration.

Examples of Confounding in Comparisons of Production Systems

laying hens

In a meta-analysis of 6,040 commercial laying hen flocks across 16 countries, which examined mortality rates in conventional cages, furnished cages, and cage-free aviaries, we found that when the maturity of the system was controlled (i.e., the level of experience with the system), there were no significant differences in mortality between caged and cage-free systems. This finding challenges the previously widespread belief that cage-free systems inherently have higher mortality rates. Instead, our analysis demonstrated that the learning curve associated with managing these systems plays a crucial role in determining welfare outcomes.

salmon aquaculture

Developmental history significantly confounds welfare comparisons across production systems. A clear example appears in salmon aquaculture, where fish reared in controlled Recirculating Aquaculture Systems (RAS) develop fundamental physiological and behavioral deficiencies that compromise their resilience in natural environments, where environmental variation and pathogen exposure are greater. When RAS-reared salmon are transferred to sea net pens, they can experience severe welfare outcomes including osmotic stress, immune suppression, and high mortality rates—not because sea pens are inherently worse for welfare, but because these fish are physiologically underprepared for marine environments. Their development in highly controlled conditions has effectively prevented the formation of crucial biological coping mechanisms necessary for survival in natural settings. This represents not merely a system quality difference but a permanent developmental limitation. Any valid comparison between RAS and sea net pen welfare outcomes must recognize that apparent system differences often reflect not just adaptation failures but irreversible developmental impairments created by early rearing conditions.

beef cattle

Feedlot vs. Pasture-Based Systems: some comparisons may suggest that cattle finished in feedlot systems experience lower disease or mortality rates than cattle in pasture-based systems. However, this does not necessarily mean that feedlot conditions are better for welfare. Feedlot cattle are more intensively monitored and more likely to receive immediate earlier veterinary interventions, which can prevent disease progression and reduce acute disease-related deaths.A proper comparison would need to control for differences in veterinary care access and breed differences to determine whether potential welfare differences are inherent to the finishing system itself or to these confounding variables.
Genetic Selection:Different genetic lines may influence the welfare outcomes observed in different production systems. For example, some high-performance beef cattle breeds used in feedlots, such as Charolais or Belgian Blue, have been intensively selected for rapid muscle growth, which can lead to increased musculoskeletal issues, heat stress susceptibility, and calving difficulties (dystocia). Conversely, breeds commonly raised in pasture systems, such as Herefords or Brahman-crosses, are often selected for heat tolerance, parasite resistance, and lower maintenance needs, which may contribute to better welfare outcomes in extensive systems. If individuals appear healthier in one system, this could be due to their genetic traits for resilience rather than the system setting itself. Controlling for breed differences allows for a fairer assessment of the welfare impacts of each system.

PIGS

Mortality Rates in Farrowing Crates vs. Loose-Housing Systems: Some studies suggest that piglet mortality is lower in farrowing crates compared to loose-housing or outdoor systems, leading to the assumption that farrowing crates provide better welfare. However, this interpretation overlooks the welfare trade-offs for both the sow and piglets. Farrowing crates severely restrict the sow’s movement, preventing natural maternal behaviors such as nest-building and bonding with piglets, critical for sow welfare. Piglets raised by less-stressed mothers and with greater opportunities for behavioral expression may also benefit from greater resilience later in life, social coping ability and immune competence, leading to overall better welfare. This is also in line with research showing that improving maternal welfare improves disease resistance, and resilience of piglets.

ANALYTICAL APPROACHES ADDRESSING COUNFOUNDING

Addressing confounding factors is crucial in animal welfare research to ensure meaningful results. Several approaches can be employed:

Best-vs.-Best Comparisons: Compare each system under optimal conditions with equivalent resource investment to highlight the true potential welfare outcomes of each system.
Worse-vs.-Worst Comparisons: Examine systems under minimal resource investment to understand their resilience to neglect or poor management.
Distribution Analysis: Evaluate the full spectrum of welfare outcomes within each system, identifying whether one system consistently produces better outcomes or merely reduces worst-case scenarios.
Economic Equivalence Approach: Compare systems at equivalent financial investment levels to assess what level of welfare can be achieved within real-world economic constraints.
Multivariate Regression Analysis: Use statistical models to adjust for multiple confounding variables simultaneously, allowing for the isolation of the effect of the primary variable of interest
Propensity Score Matching: Match subjects based on the probability of receiving a treatment given their covariates, aiming to reduce bias due to confounding variables.
Randomization: Randomly assign subjects to treatment or control groups to ensure that confounding variables are equally distributed, thereby minimizing their impact
Blocking: Arrange experimental units into groups (blocks) that are similar based on certain variables to control for variability and reduce potential confounding.
Negative Controls: Use negative control groups or outcomes to detect the presence of confounding variables and assess the specificity of the observed associations

CONCLUSION

Within the Welfare Footprint Framework, the development of a proper veterinary inventory plays a crucial role in identifying the Biological Outcomes that arise from various Circumstances animals experience. However, establishing causal relationships between these Circumstances and the Biological Outcomes identified is challenging due to multiple confounding variables—elements that can obscure or distort true causal links. Recognizing this, we place special emphasis on controlling for confounders.

Animal welfare research must adopt transparent analytical approaches to ensure robust conclusions. Without such rigor, comparisons between production systems risk being biased by resource disparities, genetic differences, lifespan variations, differences in experience with the systems and environmental factors, or mismatches of rearing and growing conditions, leading to misleading conclusions. This review serves as a reminder that carefully accounting for these variables is not just an academic necessity but a practical imperative. By employing well-controlled comparisons, we can attain a more accurate and meaningful understanding of animal welfare impacts across different production systems.

Borrowing the Knowns and Unknowns Framework to Reveal Blind Spots in Animal Welfare Science

Wladimir Alonso — Thu, 13 Feb 2025 22:45:27 +0000

Borrowing the Knowns and Unknowns Framework to Reveal Blind Spots in Animal Welfare Science

Wladimir J Alonso, Cynthia Schuck-Paim

February 13, 2025
10:45 pm

There is always a risk of fixating on what we already know while overlooking aspects of reality that might be equally or more important. Even science is not immune to this issue: often, research efforts—spanning years of publications and theses—make little meaningful progress because they remain confined to a “comfortable zone,” focusing on what is convenient to study rather than what is truly needed.

Donald Rumsfeld famously introduced the Knowns Unknowns framework for categorizing knowledge during his tenure as U.S. Secretary of Defense, specifically in the context of addressing uncertainty about whether Iraq possessed weapons of mass destruction. Its roots trace back to earlier ideas in philosophy, psychology, and risk management: for instance, the Socratic concept of ignorance, which highlights the importance of acknowledging what we don’t know, and the Johari Window (1955), which categorizes knowledge about ourselves into what is known and unknown to oneself and others. In military strategy and decision-making, the framework assists in pinpointing both anticipated and unforeseen risks.

We believe this approach is equally valuable for animal welfare science. To counter this, one of the key components of the Welfare Footprint Framework is identifying putatively neglected sources of pain before describing and quantifying them. In this context, the Knowns and Unknowns framework serves as a valuable reminder to continuously question and challenge what we think we know. We share this perspective because we believe that each quadrant of this framework can be a useful tool for others studying or working to improve animal welfare.

Figure 1. The Knowns and Unknowns Framework organizes knowledge into four quadrants: Known Knowns, where information is well-documented; Known Unknowns, where we acknowledge there are gaps in our understanding; Unknown Knowns, where knowledge exists but is not fully recognized or applied; and Unknown Unknowns, representing areas where we are unaware of what we don’t know

1. Known Knowns: What We Know We Know

In animal welfare, Known Knowns are the well-established issues that have been thoroughly researched and widely acknowledged. These are problems where the causes, consequences, and solutions are clear.

A prime example is lameness in cattle, which has been extensively studied. The circumstances causing lameness—such as poor housing conditions, inadequate hoof care, and inappropriate flooring—are well understood. Its Biological Outcomes, including chronic pain, reduced mobility, and decreased productivity, are similarly well-documented. Furthermore, solutions such as better housing designs, regular hoof trimming, and improved flooring are widely recognized as effective interventions.

The challenge for Known Knowns lies in ensuring consistent application of solutions across production systems. While the science is settled in many cases, practical implementation often lags due to economic or logistical constraints.

Figure 2. This diagram illustrates the Knowns and Unknowns Framework applied to animal welfare science, with each quadrant representing a different category of knowledge: Known Knowns, where we have a clear understanding of issues like cattle lameness, depicted by a cow struggling to walk; Known Unknowns, where we acknowledge gaps in understanding, like the welfare of salmon; Unknown Knowns, where existing knowledge such as the brain’s role in emotion is not fully utilized, shown by a mouse brain’s sagittal plane; and Unknown Unknowns, symbolized by a door, suggesting hidden or yet-to-be-discovered welfare concerns. This diagram serves as both a conceptual map and a call to action for ongoing research and application in improving animal welfare.

2. Known Unknowns: What We Know We Don’t Know

Known Unknowns in animal welfare refer to problems that are recognized but not yet fully understood or addressed. These issues highlight critical gaps in data or knowledge that must be addressed to develop effective solutions.

For example, many of the circumstances that trigger different affective states in fish are Known Unknowns. While research has established that fish can experience physical pain and other negative affective states, including fear and anxiety, our understanding on how various circumstances—such as poor water quality, crowding, or sudden environmental changes— ultimately translate into negative affective experiences remains limited. Specifically, the relationship between the perception of these circumstances by fish and the physiological and affective consequences they endure is not yet fully mapped.

3. Unknown Knowns: What We Don’t Realize We Already Know

Unknown Knowns represent knowledge that exists but has not been fully integrated into animal welfare science or applied in practice. These gaps often arise due to insufficient communication between researchers in different fields and, ultimately, scientists and animal advocates.

One particularly striking example is the neurological basis of affective states in animals. Pioneering neuroscientists such as Jaak Panksepp and Mark Solms have conclusively demonstrated that emotions like fear, joy, and pain originate from specific areas of the brain—particularly the primitive brainstem and midbrain structures—which are shared across a wide range of species. This robust body of evidence provides a clear foundation for understanding the emotional capacities of animals, yet it has not been fully embraced or utilized to comprehensively map these capacities in welfare assessments.

This gap in integration has significant implications for animal welfare. The ability to compare the capacity to experience affective states across species is fundamental to determining the moral weight of their suffering. It is also an important element for calculating the Welfare Footprint of animal products when interspecific considerations are involved. For example, weighing the welfare impacts of farming chickens versus salmon requires a clear understanding of potential differences in their emotional capacities (should one hour of Disabling pain have the same moral weight in both species? Can those groups of animals actually experience pain at those levels?).

Recognizing the importance of this issue, we at the Welfare Footprint Project are actively examining how this knowledge can be applied to interspecific welfare comparisons. This will be a critical component of future contributions, where we aim to provide practical methodologies for incorporating these insights into welfare assessments. Fully integrating the neurological basis of affective states into welfare science should facilitate more comprehensive and ethically grounded approaches to improving animal well-being.

4. Unknown Unknowns: What We Don’t Know We Don’t Know

Unknown Unknowns are the greatest blind spots. In animal welfare science, these represent welfare issues that have not yet been detected or understood, often due to limitations in current methodologies, gaps in data collection, or the inherent complexity of biological systems. By their very nature, these issues are difficult to predict or define, but they hold the potential to profoundly impact the welfare of animals.

While the specifics of these Unknown Unknowns remain speculative, parallels from human medicine might provide valuable insights. In humans, countless illnesses and conditions manifest subtly or remain undiagnosed due to diffuse and variable symptoms or the lack of adequate diagnostic tools. It is reasonable to assume that many conditions and forms of suffering in animal populations remain similarly hidden or inaccessible to current methods of observation and analysis.

It is interesting to note that this is an area where artificial intelligence (AI) may be of great value. AI excels at analyzing vast and complex datasets, uncovering patterns and anomalies that might otherwise escape human observation. By integrating data from multiple welfare indicators—such as behavioral changes, physiological markers, and environmental risks—AI can detect early warning signs of poor welfare and identify previously unknown patterns of suffering.

For example:

Subtle Behavioral Indicators: AI can analyze video footage to detect micro-behaviors that suggest stress or discomfort, even before these behaviors escalate into more obvious manifestations.
Unseen Physiological Trends: By processing biometric data, AI can identify trends and correlations indicative of pain or systemic illnesses that might not be visible through traditional monitoring methods.
Unforeseen Risks in Farming Systems: AI can model and simulate the long-term effects of farming innovations or environmental changes, predicting potential welfare risks before they emerge.

An Important Way to Address Unknowns: Transparency About the Limitations of Results

It is likely that the most significant sources of welfare loss in the most extensively studied farm species have already been identified. These include chronic pain caused by ailments associated with rapid growth and excessive productivity, chronic social stress from overcrowding, and painful routine procedures like tail docking, surgical castration, and beak trimming. However, even in these well-researched areas, we must remain vigilant. Surprising findings continue to emerge. For instance, in our analysis of broiler chickens, we initially assumed that their primary welfare burden, for individual chickens, would stem from physical ailments like lameness. However, we found that chronic hunger among breeder chickens—who are deliberately underfed to control growth—was actually the most significant source of welfare loss.

In pursuing the main goal of the Welfare Footprint Project, how do we account for the limitations of what we don’t yet know? When quantifying the Welfare Footprint of specific practices, production systems, or animal-based products, how can we ensure that potentially critical welfare aspects are not unintentionally overlooked?

Our solution is radical transparency regarding the scope and limitations of each analysis. Given the complexity of animal production systems, it is crucial to explicitly declare which key aspects—such as Life-Fates, Life-Phases, and Affective States—were included in the Welfare Footprint analysis. By doing so, we also make it clear what has been left out, ensuring that any potential gaps in knowledge are visible, rather than implicit or hidden.

For example, in our notation system (which we will detail in a future post), the Welfare Footprint of an egg will indicate:

Which Life-Fates were considered (e.g., laying hens, parent stock, and male chicks).
Which Life-Phases were included (e.g., from birth through the end of life).
Which experiences of pain and pleasure were accounted for.

By making these factors explicit, we enable meaningful comparisons between results and ensure that the scope of each analysis is clearly defined. Just as importantly, this approach creates transparency about what was left out, allowing researchers, policymakers, and advocates to recognize these Unknowns and work toward addressing them in future assessments. In doing so, we strengthen the scientific and ethical foundation of welfare evaluations, ensuring that no source of suffering goes unnoticed simply because it was not yet considered.

Using the Welfare Footprint Framework to document animal welfare in legal settings

editar — Tue, 17 Dec 2024 18:49:21 +0000

Using the Welfare Footprint Framework to Document Animal Welfare in Legal Settings

December 17, 2024
6:49 pm

The Welfare Footprint Framework (WFF) was originally developed to quantify and compare animal welfare impacts across different settings and production systems. However, its structured approach to documenting pain intensity and duration makes it particularly valuable for another critical application: providing courts with systematic, evidence-based documentation of animal suffering in cruelty cases.

Many legal systems require demonstrating that animals experienced prolonged and severe suffering to establish criminal liability in cruelty cases. This can be challenging, as assessments of suffering have traditionally relied heavily on expert opinion, which may vary between evaluators and be viewed as subjective in legal settings.

The WFF addresses this challenge through its systematic approach to documenting evidence of suffering. By breaking down experiences into discrete time segments and clearly defining different intensities of pain, the framework enables a structured evaluation of: (1) The duration of different negative experiences; (2) The intensity of suffering during each phase; (3) The scientific evidence supporting these assessments; (4) The cumulative impact on the animal over time

For example, in cases involving severe neglect, the framework can help document (1) How long animals went without food or water; (2) The intensity of their suffering during different phases (of starvation/dehydration); (3) scientific evidence of their distress (neurological, physiological, behavioral, pharmacological, etc), (4) The cumulative time spent in different states of suffering.

Rather than relying solely on expert opinion, the WFF requires documenting specific evidence that supports or contradicts each assessment of pain intensity. This evidence-based approach helps demonstrate to courts that assessments of suffering are grounded in scientific findings rather than subjective judgment.

The framework’s potential in legal settings was recently explored in a workshop with Costa Rica’s National Animal Health Service (SENASA) and Judicial Investigation Agency (OIJ). Over two days, professionals from both agencies examined how the WFF could be applied to strengthen documentation of animal suffering in legal proceedings. The structured nature of the framework resonated with participants as a way to bridge the gap between scientific evidence and legal requirements in animal welfare cases.

While originally designed for different purposes, the framework’s systematic approach to quantifying animal welfare experiences makes it a valuable tool for demonstrating suffering in legal settings. The interest from agencies like SENASA and OIJ suggests promising potential for its application in animal cruelty cases.

Our Scientific Director Discusses WFP’s Methodology on Hear This Idea

editar — Thu, 21 Nov 2024 15:46:42 +0000

Quantifying Animal Welfare: Our Scientific Director Discusses WFP's Methodology on Hear This Idea

November 21, 2024
3:46 pm

Our Scientific Director, Cynthia, recently joined Fin Moorhouse on Episode 81 of the Hear This Idea podcast to discuss the Welfare Footprint Project’s innovative approach to quantifying animal welfare. The conversation offered a comprehensive look at how we’re working to create standardized, evidence-based metrics for measuring animal experiences across different species and farming systems.

A Cross-Species Framework for Measuring Welfare

One of the central challenges in animal welfare assessment has been developing a methodology that allows for meaningful comparisons across different species and farming conditions. During the interview, Cynthia explains our approach using Cumulative Pain and Cumulative Pleasure metrics – a system that measures negative and positive experiences across four intensity levels. This framework considers both the intensity and duration of experiences, allowing us to assess cumulative welfare impacts throughout any period or an animal’s whole life.

Beyond Traditional Welfare Metrics

Traditional welfare assessment methods often rely on species-specific scoring systems that make cross-comparisons difficult. Our method takes a different approach, examining fundamental aspects of animal experiences that can be measured across species. By looking at indicators such as behavior, physiological responses, pharmacology and evolutionary biology, we can estimate welfare impacts with a transparent and evidence-based approach.

The Real-World Impact

The podcast discussion delves into practical applications of our methodology, from assessing laying hen welfare in different housing systems to evaluating welfare impacts in salmon farming. We also discuss how faster-growing breeds in modern farming affect animal welfare.

Looking Forward: Technology and Transparency

Cynthia also discusses exciting developments in our work, including:

The potential of AI tools to help scale welfare assessments
Whether the ability to feel pain is unique to big brained animals, or more widespread in the tree of life
How fish farming compares to poultry and livestock farming
Whether positive experiences like joy could make life worth living for some farmed animals
How animal welfare advocates can learn from anti-corruption nonprofits
Our upcoming project focusing on the Welfare Footprint of the Egg

This episode offers valuable insights into our methodology and vision for the future.
Listen to the full conversation here to learn more about how we’re working to create more evidence-based approaches to animal welfare assessment and improvement.

If you don’t know the ‘Hear This Idea‘ podcast yet, I highly recommend checking it out. Each episode dives deep into fascinating and thought-provoking conversations with researchers and practitioners working on some of the world’s most pressing problems. The hosts have a remarkable talent for making complex ideas accessible while preserving their rigour and depth. Take a look at their other episodes – they’re well worth your time!

Cynthia recording the Hear This Idea podcast from a makeshift workspace while traveling…

Bridging Neuroscience and Philosophy: Exploring Animal Emotions and Welfare with the Neurophilosopher GPT Tool

editar — Fri, 13 Sep 2024 14:30:36 +0000

Bridging Neuroscience and Philosophy: Exploring Animal Emotions with the Neurophilosopher GPT Tool

Wladimir J Alonso, Cynthia Schuck-Paim

September 13, 2024
2:30 pm

Understanding where emotions like love, happiness, pain, and suffering arise in the brain is one of the most profound inquiries we can explore. For centuries, this topic was considered beyond the scope of scientific investigation. Today, however, neuroscience has made significant strides in uncovering the origins of these emotional experiences. This growing knowledge not only deepens our understanding of human nature but also opens doors to practical applications, such as assessing how different species experience pain—critical for measuring animal welfare.

Despite these advances, neuroscience can still seem intimidating, limiting the spread of these insights. But it doesn’t have to be this way. Philosophy, an unexpected but powerful ally, offers a unique lens that complements and clarifies neuroscientific knowledge in exploring these questions. By integrating philosophy with neuroscience, we can make complex ideas more accessible and engaging. In fact, many philosophical dilemmas are rooted in conflicts between different brain systems—age-old debates that have been discussed by thinkers like Aristotle, Kant, Nietzsche, Freud, and Descartes. By revisiting these debates from a neuroscientific perspective, we can make the journey to understanding animal brains and minds, including our own, far more enlightening.

This is where neurophilosophy comes in—a field that bridges neuroscience with philosophical questions about the mind, consciousness, and cognition. Foundational thinkers like Patricia Churchland have shown how integrating these fields can reshape traditional debates about our mental and emotional lives.

Bridging Neuroscience and Philosophy: The Old Brain vs. New Brain

One effective way to grasp the complexity of the brain is through its evolutionary development. In simplified terms, the brain can be divided into two systems: the ‘old brain’ and the ‘new brain.’ The ‘old brain’ includes ancient structures responsible for survival functions and basic emotions like fear, hunger, and instincts, governed by areas such as the brainstem and limbic system. The ‘new brain,’ centered in areas such as the neocortex in mammals, is responsible for higher-order cognitive functions like abstract thinking, decision-making, and moral reasoning.

As the brain evolved, these two systems were layered upon one another. Today, they coexist—sometimes harmoniously, but often in conflict. This tension is where many philosophical and ethical dilemmas arise, as our primitive emotional desires frequently clash with our higher-order reasoning, shaped by morality, social norms, and long-term goals. The interaction between these brain systems is central to neuroscience and provides a valuable framework for exploring key philosophical questions about behavior, decision-making, and values.

Introducing the Neurophilosopher GPT Tool

Traditionally, the study of neuroscience, philosophy, and neurophilosophy has been reserved for scholars or those with specialized backgrounds. With the Neurophilosopher GPT tool, however, we’ve created an interactive platform where anyone can explore these fields —whether you’re a complete beginner or an experienced scholar. By inputting any philosophical author, theory, moral question, or religious idea, you can engage in dynamic discussions that provide detailed neuroscientific analysis through the ‘old brain’ and ‘new brain’ framework.

This tool transforms passive learning into an active, personalized experience. It enables exploring the scientific and philosophical foundations of emotions, consciousness, and thought in a way that makes these subjects accessible and relevant to your everyday life.

The tool draws heavily from a branch of neuroscience called affective neuroscience, which focuses on how emotions (or affective states) arise in the brain. Neuroscientists like Jaak Panksepp, Antonio Damasio, and Mark Solms have argued that emotional experiences are ancient and shared across many species, deeply rooted in the older structures of the brain. In contrast, cognitive neuroscientists often emphasize that conscious emotions are tied to the more advanced neocortical areas found in humans and their closer relatives. For those interested in this ongoing debate, we recommend exploring the detailed discussions between proponents of these perspectives in this paper.

Therefore, while Affective Neurophilosophy would have more accurately described this tool’s focus, we chose to stick with Neurophilosophy to keep it approachable. The term already bridges two complex fields—neuroscience and philosophy—and we didn’t want to make it seem more intimidating. Our goal is to make the tool as inviting and accessible as possible for anyone to explore.

Practical Applications and Exploration

The tool also encourages reflection on contemporary issues with significant ethical implications, such as what brain science can reveal about the capacities for pain and pleasure in other species. If emotional experiences are deeply rooted in ancient brain structures (the ‘old brain’), this raises crucial ethical questions about the extent to which different species, based on their neural architecture, should receive special consideration in terms of suffering prevention and mitigation. These inquiries not only deepen our understanding of our own brains but also guide broader ethical considerations.

Whether you’re interested in exploring classic philosophical ideas or investigating the brain’s role in moral and ethical questions, the Neurophilosopher GPT tool is helpful. You can use the tool to examine how Aristotle’s views on the soul, John Stuart Mill’s distinction between higher and lower pleasures, Jeremy Bentham’s views on animal suffering, Peter Singer’s arguments for animal liberation, Thomas Nagel’s exploration of consciousness, or René Descartes’ perspectives on animal consciousness align with neurophilosophy. The possibilities are vast, providing fresh insights into how our ancient and modern brain systems shape emotions, thoughts and decisions.

We invite you to try the Neurophilosopher GPT tool and discover how the intersection of philosophy and neuroscience can provide a deeper understanding of the capacities for pain and suffering across species, enriching our perspectives on the welfare of animals.

SOME RECOMMENDED READING SOURCES

Churchland, P. S. (1986). Neurophilosophy: Toward a Unified Science of the Mind-Brain. Cambridge, MA: MIT Press.
Churchland, P. S. (2002). Brain-Wise: Studies in Neurophilosophy. Cambridge, MA: MIT Press.
Churchland, P. M. (1984). Matter and Consciousness: A Contemporary Introduction to the Philosophy of Mind. Cambridge, MA: MIT Press.
Churchland, P. M. (1995). The Engine of Reason, the Seat of the Soul: A Philosophical Journey into the Brain. Cambridge, MA: MIT Press.
Damasio, A. R. (1994). Descartes’ Error: Emotion, Reason, and the Human Brain. New York: G.P. Putnam’s Sons.
Damasio, A. R. (1999). The Feeling of What Happens: Body and Emotion in the Making of Consciousness. New York: Harcourt Brace.
Dennett, Daniel C. (1991). Consciousness Explained. Little, Brown and Company.
Gazzaniga, Michael S. (2018). The Consciousness Instinct: Unraveling the Mystery of How the Brain Makes the Mind. Farrar, Straus and Giroux.
LeDoux, Joseph E. (1996). The Emotional Brain: The Mysterious Underpinnings of Emotional Life. Simon & Schuster.
MacLean, Paul D. (1990). The Triune Brain in Evolution: Role in Paleocerebral Functions. Springer.
Panksepp, J. (1998). Affective Neuroscience: The Foundations of Human and Animal Emotions. New York: Oxford University Press.
Panksepp, Jaak, Richard D. Lane, Mark Solms, and Ryan Smith. (2017). “Reconciling Cognitive and Affective Neuroscience Perspectives on the Brain Basis of Emotional Experience.” Neuroscience and Biobehavioral Reviews 76 (Pt B): 187–215.
Solms, Mark. (2019). “A Neuropsychoanalytic Perspective on the Hard Problem of Consciousness (I).” בטיפולנט קהילה מקצועית; Youtube. February 24, 2019. YouTube Link.
Solms, Mark. (2021). The Hidden Spring: A Journey to the Source of Consciousness. W. W. Norton & Company.

Consulting Opportunity: Beef Cattle Welfare Research

editar — Sun, 01 Sep 2024 09:07:53 +0000

Opportunity: Consulting Work on Beef Cattle Welfare

The Welfare Footprint Project is a scientific initiative aimed at quantifying animal welfare impacts using a meaningful, relatable, and comparable metric: time spent in affective states (negative or positive) of varying intensities.

We are currently seeking a beef cattle welfare specialist to work on a project quantifying the welfare impacts of several interventions designed to promote higher welfare standards in beef cattle production in South America. This consulting service will involve using the Welfare Footprint framework for impact assessment.

Key Responsibilities

Literature review and data analysis: review and analyze data related to beef cattle welfare from various production environments.
Welfare Assessment: apply the Welfare Footprint framework to evaluate the welfare impact of different interventions.
Reporting: write detailed reports and presentations summarizing findings, including the estimated welfare impacts and areas for improvement.

who we are looking for

We are seeking researchers who have:

Knowledge of the beef cattle production chain in Latin America.
Knowlege of beef cattle welfare and welfare indicators (e.g., physiology, neurology, pharmacology, behavior).
Experience conducting literature reviews.
Analytical and critical skills, and proactive approach to problems (‘can-do’ attitude)
Ability to work independently and meet deadlines
Great level of attention to detail
Fluency in English
Knowledge of Portuguese desirable

CONTRACT DETAILS

Location: Remote
Compensation: BRL 10,000/month (or equivalent in another currency)
Hours: 36 hours/week
Work to start on Oct 07

APPLICATION

Applications are now closed.

Can AI power the Global Mapping and Quantification of Animal Suffering? The Pain Atlas Project

editar — Tue, 25 Jun 2024 17:46:38 +0000

Can AI power the Global Mapping and Quantification of Animal Suffering? the Pain Atlas Project

Wladimir J Alonso, Cynthia Schuck-Paim

June 25, 2024
5:46 pm

Demis Hassabis is one of the key figures behind many technological advancements in AI. In his TED talk “How AI Is Unlocking the Secrets of Nature”, he illustrates a groundbreaking achievement where AI saved an estimated one billion year’s worth of research hours in molecular biology. Traditionally, solving the three-dimensional structure of a protein takes a PhD student 4-5 years, but with AI this could be done in minutes. In just one year, the AI-based AlphaFold project described the structure of the 200 million known proteins, which will greatly contribute to advance our understanding and treatment of many diseases. The capabilities of AI have the potential to revolutionize other scientific fields too, at a pace that is difficult to comprehend.

We propose that one such field is the large-scale mapping and quantification of animal suffering, and that the analytical tools required for this task are already available.

Recruiting AI to help quantify animal suffering

AlphaFold used AI by training a deep learning model on a vast dataset of known protein structures and their corresponding amino acid sequences, enabling it to predict new structures from new sequences. To quantify (negative or positive) affective experiences, we depart from a very different reality, since there is not yet any method that enables directly accessing subjective affective states. Therefore, different from proteins, there is no dataset of known correspondences that can be used by deep learning models.

However, a method to estimate the key elements of affective experiences – duration and intensity – has been recently made available through the development of the Welfare Footprint Framework (WFF). Over the last few years, the WFF has been used to estimate the welfare impact of animal welfare reforms and interventions in different species. Although originally developed for use by researchers, the WFF is nevertheless highly suitable for use by AI, specifically in Large Language Models (LLMs) like OpenAI’s GPT series, Google’s Gemini, Anthropic’s Claude, and Meta’s LLaMA.

One particularly critical part of this process, where the WFF made its major contribution, is in the description and quantification of affective states. This is explained in more detail elsewhere, but in short, the method for describing negative experiences is performed with the Pain-Track (Fig. 2) and follows two basic stages:

Breaking Down the Experience: because the intensity of most experiences fluctuates over time (e.g., it may increase as a disease advances, or reduce as healing progresses), each experience is divided into meaningful time segments, each representing an expected intensity pattern. For example, the pain from stubbing a toe starts with sharp, immediate pain lasting seconds, followed by acute pain with swelling, lasting minutes. This continues through secondary pain and bruising (minutes to hours), subacute pain and healing (hours to days), and mild residual pain during recovery. This approach can be applied to any unpleasant experience, such as disease, injury, fear, frustration, or hunger.

Estimating Intensity and Duration during each Segment: for each temporal segment, the intensity of the negative experience is estimated, ranging from Annoying to Excruciating. To transform scattered knowledge from various fields into estimates within each cell in the Pain-Track, the WFF uses a structured approach that documents existing evidence and highlights how well it supports or contradicts each intensity hypothesis. For example, to estimate the intensity of pain a make piglet endures during surgical castration, a procedure whereby piglets are restrained, the scrotum incised with a scalpel, and the testes extracted by tearing the tissues or severing the spermatic cord, the experience is first divided into meaningful time segments. Next, all relevant information is gathered from the literature. This includes, among others, data on nociceptor densities in this tissue, the animal’s behavior, posture and vocalizations during and after the process, neurological and physiological responses, the characterization and typical duration of inflammatory processes, the type and dose of analgesics required to alleviate pain, and the evolutionary significance of pain for the animal and in the area affected. Each of these lines of evidence is then compared with the definitions of the four categories of pain intensity used in the WFF to estimate the intensity and duration of the experience in each of these specific moments. Once estimates of the intensity and duration of the experience are made for each time segment of a Pain-Track, calculating welfare loss, measured as Cumulative Time in Pain of different intensities (or Cumulative Pain for short), is straightforward and automatic.

Figure 2. Pain-Track with hypotheses about the temporal evolution of pain intensity in piglets castrated without pain relief, and resulting cumulative time in pain, of each intensity (Cumulative Pain). Estimates of intensity and duration are based on a comprehensive review of evidence (not shown here) on indicators of pain (behavior, physiology, neurology, pharmacology, immunology, evolutionary reasoning) at each temporal stage. Cumulative pain estimates discount eight hours per day from sleeping.

Since discovering the powerful capabilities of LLMs to gather and interpret large volumes of data, we have been exploring their potential, particularly in creating Pain-Tracks. One outcome of this exploration is the ‘Hedonic-Track’ Custom GPT. This tool provides a starting point for describing and quantifying the impact of various welfare issues across species. In the video below, you can see an example of its operation, where the user simply confirmed each step by answering ‘yes’. However, interactions can be much more detailed. For instance, PDFs with relevant information can be uploaded during the conversation, answers can be challenged at any stage, and further recommendations can be requested on the topic, among other possibilities.

Video demonstrating how the Hedonic-Track Custom GPT (former 'Pain-Track Custom GPT') operates at the time of this writing, using the welfare impact of air asphyxia in fish (trout) as an example.

From the moment the tool has started to perform satisfactorily, our priority has been to make it available to the research and advocacy community, even if they are not yet perfect. We found it particularly useful for didactic purposes, incorporating it into our workshops as a way to engage participants in the understanding and use of the Welfare Footprint method. Participants are motivated by being able to immediately describe and quantify sources of suffering in their target species or even for their personal pain experiences.

While the results of the tool should only be interpreted as a starting point of analysis, which still requires human revision, estimates of Cumulative Pain produced so far are promising. For example, except for Annoying pain, estimates of Cumulative Pain due to surgical castration in male piglets produced by the Hedonic-Track Custom GPT tool fall within the credibility interval of estimates developed by researchers, as shown below:

Still, LLMs are not deterministic, so estimates vary across queries. In the example above, 3 out of 10 individual estimates of Excruciating pain, and 5 (out of 10) individual estimates of Disabling pain did not overlap with those produced by the researchers, despite being in the same ‘ballpark’. Therefore, if the outputs of this GPT tool are aimed at informing decision-making, we advise using the average of about 10 queries, in the same way that human-based estimates are typically the consensus of various researchers.

The Pain ATLAS PROJECT

Identifying and understanding the sources of suffering (mainly in its extreme forms) in humans and animals under our custody are arguably among the most morally important research endeavors we can embark on. Given that the technological conditions are ripe, as we have shown throughout this text, we propose the ‘Pain Atlas Project’ to make this possible. By potentially utilizing higher levels of AI resources or collaborating with an AI company or institution, we hope to achieve an advance in animal welfare sciences comparable to what AlphaFold accomplished for molecular biology.

This project is designed to be structured around three core components:

1. Mapping of Suffering

This component involves a comprehensive analysis of the primary source of suffering endured by different animal species throughout their lives and various contexts (Figure 1, sections I & II). This includes conditions such as injuries, diseases, deprivations, and stressors, building on efforts such as the Veterinary Extension of SNOMED CT.

2. Quantification of Suffering

This component involves use of the Cumulative Pain metric to estimate the magnitude of suffering associated with each of the sources of suffering identified (Figure 1, section III). In a large-scale effort such as the Atlas, the instructions would be more detailed than those allowed with user-end resources like Custom GPTs. For instance, the results for each Pain-Track should be the consensus (or average) of the results obtained from different ‘agents’, each with a different specialty (e.g., ‘physiologist’, ‘veterinary’, ‘neuroscientist’)

3. Visualization of Suffering

This final component uses visualization tools to construct a detailed and global landscape of suffering across species and living conditions. The focus will be on highlighting hotspots of suffering and assessing the effectiveness of potential interventions. This analytical phase is crucial for transforming raw data into actionable insights, ultimately guiding decision-making and intervention strategies..

We invite everyone to provide feedback in the EA forum and discuss potential collaborations (feel free to also reach out to us at AI@welfarefootprint.org).

Quantifying Positive Animal Welfare in Individuals and Populations

editar — Tue, 12 Mar 2024 19:08:00 +0000

Beyond Suffering: A Framework for Quantifying Positive Animal Welfare in Individuals and Populations

Wladimir J Alonso, Cynthia Schuck-Paim

Center for Welfare Metrics, Brazil

Main Takeaways

The Welfare Footprint Project has so far focused predominantly on quantifying negative affective experiences in animals, given pain’s greater impact on welfare.
However, positive affective states also play a crucial role in an animal’s quality of life, affecting long-term welfare through aspects such as immunity and resilience.
An operational definition of pleasure is proposed.
Categories of pleasure intensity are defined based predominantly on the degree of engagement with positive experiences.
The Pleasure-Track notation system is proposed, allowing for the description of the temporal evolution of the intensity of positive affective states.
Cumulative Pleasure is measured as the total time spent in positive affective states of different intensities.
The framework avoids equating the intensity categories of pleasure and pain, and acknowledges the complexity of balancing positive and negative affective states.
A graphic proposal is made for the visualization of Welfare Footprints, where they are expressed by a double bar chart representing the time spent in negative and positive affective states of different intensities (Cumulative Pain and Cumulative Pleasure).
animal experiences pain

INTRODUCTION

To inform efforts for the prevention and alleviation of animal suffering, the Welfare Footprint Project has so far primarily focused on quantifying negative affective experiences [1], often referred to simply as ‘pain.’ This emphasis is based on the understanding that negative experiences generally have a greater impact on well-being than positive ones, often preventing the experience of positive states [2]. From an evolutionary standpoint, pain is expected to be more noticeable than pleasure, as pain signals immediate threats to survival and reproduction, whereas positive experiences tend to reinforce behaviors that are beneficial in the medium term, like social bonding, learning, optimal foraging and mating. Failing to experience pleasure is not as immediately consequential as not responding to pain. This tendency to focus more on negative than positive events, known as negativity bias, is in fact a well-documented phenomenon in humans [3].

Still, positive states also play a crucial role in shaping an individual’s quality of life, and determining the extent to which it is ‘good’ [4], or at least ‘worth living’ [5]. Positive states are not only momentarily pleasurable; they also have long-term effects that help individuals overcome adversity [6]. For example, experiences of positive affect have been shown to naturally relieve pain (through endogenous analgesia), boost resilience to stress, and improve immune function [7–9]. Accordingly, the assessment of positive affect and the identification of conditions promoting it have gained substantial traction in the animal welfare sciences in recent years [10–12].

The goal of this contribution is to expand the Welfare Footprint framework to include not only the measurement of negative experiences in animals, but also positive states, or ‘pleasure’. Since the reasoning behind this metric is similar to that for assessing pain — already discussed in prior works [1,13] (see a 8 min video or a presentation at the Effective Altruism conference) — we do not describe these principles again here. Instead, we introduce the definitions and tools equivalent to those used for the quantification of negative affective states, now adapted for positive affect: the operational definitions of ‘pleasure’ and its four different intensity levels, the ‘Pleasure-Track’ notation system, and the ‘Cumulative Pleasure’. Additionally, with the full spectrum of affective experiences considered, we also introduce a visualization proposal, termed ‘Cumulative Affect’, to represent the overall Welfare Footprint of any population in a context and time scope of interest.

Operational definition of pleasure

We propose an operational definition of pleasure directly derived from the operational definition of pain previously proposed [14], as follows:

“Pleasure is a conscious experience, evolved to elicit or reinforce behaviors beneficial to an organism’s survival and/or reproduction. It is affectively and cognitively processed as a positive and dynamic sensation that can vary in intensity, duration, texture, spatial specificity, and anatomical location. Pleasure is characterized as ‘physical’ when primarily triggered by stimuli that are directly rewarding or enjoyable, and as ‘psychological’ when triggered by cognitive processes, memories, and emotional states. Depending on its intensity and duration, pleasure can override other adaptive instincts and motivational drives, leading to states of dependency and self-damage.”

Categories of Pleasure Intensity

In the Welfare Footprint Framework, pain intensity categories are operationally defined based on the assumption that more unpleasant sensations should be more disruptive, engaging a greater share of attention. [15,16]. This is rooted in the evolutionary expectation that the greater the threat, the more intense the signal should be to ensure it is prioritized over other functions and behaviors [17,18]. A similar approach can be used to define categories of pleasure intensity, with intensity categories defined based on the degree of engagement with positive experiences. First, the degree of engagement in experiences is likely to correspond to the hedonic value of these experiences [6]. A greater motivation to play, interact socially and explore is likely driven by a more intense positive experience. Second, the degree of motivation to engage in positive experiences is also likely to match their broader adaptive value [19]. Just as pain intensity often signals threat severity, pleasure intensity may correlate with the evolutionary importance of the activity, such as resource holding, learning, parental care, and mating (though maladaptive exceptions can be present, such as addiction in humans and feather-plucking in birds).

Similarly with the case of pain, here we propose the use of emphatic and universally recognized terms, focusing on concepts that reflect the degree of engagement with the pleasurable state. The definitions of pain intensities were first presented in a paper aimed at the medical audience [13], but designed to maintain their universality for non-human animals. This same broad approach is used to define pleasure intensities:

Satisfaction: low-intensity positive states, where an individual shows subtle signs of comfort or satisfaction associated with a physical sensation or meeting a non-essential need. These might include comfortable bedding conditions, grooming or basking in the sun. Engagement is present but not overwhelming, allowing the individual to easily shift attention to other stimuli or activities as needed.
Joy: positive states involving greater engagement in rewarding activities. Individuals may display enhanced vigor in play, stronger social bonding, or more or active engagement in highly preferred activities, such as foraging. Behaviors indicative of joy suggest a greater focus on these positive experiences, although they do not exclusively dominate the individual’s attention. The individual’s behavior is noticeably directed towards maintaining or enhancing the positive experience. Physiological indicators may include heightened autonomic responses (e.g., heart rate).
Euphoria: experiences in this category are intense and the primary focus of attention. Everything else might seem secondary. Euphoria might be observed in immersive play, mating rituals, or the pursuit and enjoyment of highly favored resources, such as a successful hunt. In some situations, this intense state might lead to spontaneous expressions of pleasure, such as vocalizations.
Bliss: At the peak of positive experiences, bliss represents a profound level of pleasure that pervades the individual’s sensory and emotional experience. It’s a sensation that transcends the ordinary. When experiencing bliss, the sensation of pleasure is so overwhelming that it eclipses other immediate needs or environmental stimuli for the duration of the experience. The world outside fades away as the individual is consumed by this all-encompassing state. Examples could be orgasmic states, reuniting with socially significant partners that are long missing, or the encounter of extremely positive conditions after prolonged periods of stress and hardship. Blissful states are expected to be rare and of short duration.

Just as we recently included the category ‘no pain’ as an additional intensity level (zero pain), it is also convenient, whenever relevant, to include the category ‘no pleasure,’ which represents the hypothesis that no positive affective states are experienced over the period of interest.

Pleasure-Tracks

The Pleasure-Track is a notation system analogous to the Pain-Track [13], where hypotheses on the temporal evolution and intensity of positive affective states are described. To this end, each experience is atomized into a level of analysis that is justifiable through empirical evidence. This is done by decomposing each experience into meaningful time segments, where each time segment is characterized by an expected intensity. Next, hypotheses about the intensity and duration of the experience at each of these segments can be informed by empirical evidence.

To illustrate the concept of a Pleasure-Track, we consider the hypothetical intensity and duration of play bouts in young calves, as follows: (Phase I) the play session begins with calves engaging in light play or exploration. This includes behaviors such as gently nudging or sniffing play objects, or light social interactions with other calves. The engagement is present but not overwhelming, allowing the calves to remain aware of their surroundings and easily shift their focus to other stimuli. As the play behavior escalates (Phase II), calves enter a state characterized by more vigorous activities such as running, jumping, and robust social play like head-butting or chasing. The calves show a clear focus on these rewarding activities, with their behavior directed towards maintaining or enhancing the experience. Physiological indicators might include increased heart rate or more expressive body language, reflecting their engagement in the play. The peak of the play experience (Phase III) is where the activity becomes the primary focus of the calves’ attention, with calves fully immersed in play, displaying spontaneous expressions of pleasure such as vocalizations or exuberant body movements. After reaching the peak of excitement, the intensity of play begins to decrease. Calves may start to engage in less vigorous activities, such as slower running or gentle nudging, as they start to wind down from the high energy play. This stage (Phase IV) is characterized by a gradual reduction in the intensity of their actions and a shift towards more relaxed behaviors. When the play bout concludes (Phase V) calves often display behaviors indicative of relaxation. This may include lying down, social grooming, or simply resting in close proximity to their playmates. The Pleasure-Track below summarizes these hypothetical ethological observations for a group of calves. Average phase durations are also hypothetical, and in real scenarios are expected to vary with factors such as the species, age, space available, group size, climatic conditions, time of day, and hunger [20].

As with the Pain-Track, the Pleasure-Track is designed to capture situations in which there is uncertainty in the classification of the intensity of pleasure. Therefore, probabilities are used to represent how likely it is that each category of pleasure intensity is experienced. Uncertainty (or natural variation) regarding how long each phase of the experience lasts is captured in the range of values at the bottom row of each temporal segment. The Pleasure-Track illustrated above is hypothetical, but in real situations, each numerical input should be based on a thorough review of evidence from various sources. Because evidence will be often limited, criticism of the proposed values should be always encouraged.

Cumulative Pleasure (individual level)

Like with estimates of cumulative time in negative states (‘Cumulative Pain’), it is also possible to estimate ‘Cumulative Pleasure’, namely the time spent in positive affective states of different intensities, as follows:

The assessment of Cumulative Pleasure for an individual over a certain period or even their lifetime, can be also established by determining the cumulative impact of all the positive events experienced. Similar to negative experiences, this is best represented as the sum of the time spent in positive states from all these events, whether they happen sequentially or simultaneously. This also includes events that happen repeatedly. For instance, if young calves play as described in the hypothetical scenario 1-8 times a day for 4 weeks, the Cumulative Pleasure for each calf would be a total of approximately 15-130 minutes of Euphoria, 1 to 7.3 hours of Joy, and 1.7 to 13.3 hours of Satisfaction. This is the of the estimates shown in the Pleasure-Track, daily frequency of playing bout and total number of days playing.

Of course, different positive experiences can interact with each other in various ways, both physically and psychologically. These effects, when potentially leading to changes in the outcomes, must be investigated on a case-by-case basis.

Cumulative Pleasure (population level)

Cumulative Pleasure can be also calculated at the population level. This requires accounting for differences in the exposure of population members to different positive experiences. This is achieved by weighting estimates by the prevalence of each experience, which enables determining the cumulative time in pleasure experienced by an average member of the population. For example, in the hypothetical scenario of play in calves, Cumulative Pleasure could be weighted by the proportion of the calf population playing in the period of interest. For instance, if 50-90% of the calves played over the period of two weeks, Cumulative Pleasure for the average population member over this period would be the product of this prevalence and Cumulative Pleasure for each individual over the two weeks (resulting in 15-130 minutes of Euphoria, 59-440 minutes of Joy and 100-800 minutes of satisfaction)

As with estimates of time spent in pain, we don’t combine the four categories of pleasure intensity into a single one. This is because, as thoroughly discussed for pain [21], there are no empirical references to establish a weighting system between the intensity categories, hence a single scale of pleasure.

A notation proposal for Welfare Footprints

By considering positive experiences, Welfare Footprints can be expressed as the time spent in negative and positive affective states of different intensities, i.e., Cumulative Pain and Cumulative Pleasure. To help visualize these effects, we propose to present Welfare Footprints as a double bar chart, as illustrated in the examples below. This graphical representation juxtaposes the time spent in negative affective states against the time spent in positive affective states, across various intensities. The figure illustrates two footprints, depicting the distribution of time spent by an individual (or population, if the prevalence of different pains and pleasures are factored in the analysis) at each intensity of pain and pleasure (note that the quality of life depicted in the first Welfare Footprint [A] is clearly better than in the second [B]):

The time units used in the chart for each intensity category (seconds, hours, days, and months) are flexible and can be adjusted. However, because of this flexibility, it is crucial to always specify the units in the chart.

Because positive and negative intensities are displayed side by side, it is important to reiterate that no equivalence is implied. For instance, whether 10 seconds on the most intense form of pleasure (‘Bliss’) can be considered a direct counterpart of 10 seconds under the most intense form of pain (‘Excruciating’) remains to be determined.

Further Thoughts

The welfare of sentient organisms is shaped by a complex interplay of positive and negative affective states [6,19,22]. While combining both positive and negative experiences would offer a fuller picture of their affective lives, there are many challenges to measure or compare these states ethically or quantitatively [5,23–27]. For example, Shriver [26] challenges the idea that pleasure and pain are just two ends of a continuum, pointing out that these states are driven by different cognitive processes, contribute differently to overall well-being, and have separate relationships with motivational systems. Considering these complexities, we do not attempt estimating how cumulative time in pain and pleasure might balance each other out. Instead, we focus on describing and measuring these affective states in a transparent and relatable way: by estimating time spent in states of pleasure and pain, at different levels of intensity.

We still maintain that negative states have a disproportionate impact on an individual’s life and welfare. Pain, especially when severe, can prevent the possibility of positive experiences. However, positive experiences and states are valuable, especially when more intense sources of pain have already been mitigated. Therefore, Welfare Footprints must not overlook this aspect. This text is aimed at incorporating positive experiences into the Welfare Footprint framework. Welfare Footprints that focus solely on pain are already valuable (especially for addressing the extreme situations many captive animals face), but a more comprehensive approach that includes positive experiences can provide a richer assessment. This complete version of a Welfare Footprint can thus better guide animal welfare policies, inform advocacy groups, and enhance public awareness about the ethical implications of animal use.

References

1. Alonso WJ, Schuck-Paim C. The Comparative Measurement of Animal Welfare: the Cumulative Pain Framework. In: Schuck-Paim C, Alonso WJ, editors. Quantifying Pain in Laying Hens. Independently published. https://tinyurl.com/bookhens; 2021.
2. Schuck-Paim C, Alonso WJ. How attention modulates the perceived intensity and duration of simultaneous affective experiences:Implications for refining Cumulative Pain estimates and for determining the potential for positive welfare. In: Welfare Footprint Project [Internet]. 16 Apr 2023 [cited 5 May 2023]. Available: https://welfarefootprint.org/2023/04/16/method-refinement-attention/
3. Norris CJ. The negativity bias, revisited: Evidence from neuroscience measures and an individual differences approach. Soc Neurosci. 2021;16: 68–82.
4. Rowe E, Mullan S. Advancing a “Good Life” for Farm Animals: Development of Resource Tier Frameworks for On-Farm Assessment of Positive Welfare for Beef Cattle, Broiler Chicken and Pigs. Animals. 2022;12: 565.
5. Yeates J. Better to have lived and lost – the concept of a life worth living. In: Butterworth A, editor. Animal Welfare in a Changing World. CAB International; 2018. pp. 162–170.
6. Leknes S, Tracey I. A common neurobiology for pain and pleasure. Nat Rev Neurosci. 2008;9: 314–320.
7. Yeates JW, Main DCJ. Assessment of positive welfare: a review. Vet J. 2008;175: 293–300.
8. Moskowitz JT, Saslow LR. Health and psychology: The importance of positive affect. In: Tugade MM, editor. Handbook of positive emotions (pp. New York, NY, US: The Guilford Press, xv; 2014. pp. 413–431.
9. Gentle MJ. Attentional Shifts Alter Pain Perception in the Chicken. Anim Welf. 2001;10: 187–194.
10. EU. Action project to provide the background for including positive welfare in farm animal welfare assessment. In: LIFT – COST ACTION [Internet]. Super Admin; 2022-2026 [cited 28 Feb 2024]. Available: https://liftanimalwelfare.eu/
11. Mellor DJ. Animal emotions, behaviour and the promotion of positive welfare states. N Z Vet J. 2012;60: 1–8.
12. Mellor DJ. Updating Animal Welfare Thinking: Moving beyond the “Five Freedoms” towards “A Life Worth Living.” Animals (Basel). 2016;6. doi:10.3390/ani6030021
13. Alonso WJ, Schuck-Paim C. Pain-Track: a time-series approach for the description and analysis of the burden of pain. BMC Res Notes. 2021;14: 229.
14. Alonso WJ, Schuck-Paim C. A Novel Proposal for the Definition of Pain. pre-print (OSF). 2023; 2.
15. Barclay RJ, Herbert WJ, Poole T. The disturbance index : a behavioural method of assessing the severity of common laboratory procedures on rodents. UFAW, Universities Federation for Animal Welfare; 1988.
16. Eccleston C, Crombez G. Pain demands attention: a cognitive-affective model of the interruptive function of pain. Psychol Bull. 1999;125: 356–366.
17. Merker B. Drawing the line on pain. Animal Sentience: An Interdisciplinary Journal on Animal Feeling. 2016;1: 23.
18. Mellor DJ, Beausoleil NJ, Littlewood KE, McLean AN, McGreevy PD, Jones B, et al. The 2020 Five Domains Model: Including Human-Animal Interactions in Assessments of Animal Welfare. Animals (Basel). 2020;10. doi:10.3390/ani10101870
19. Mellor DJ. Positive animal welfare states and encouraging environment-focused and animal-to-animal interactive behaviours. N Z Vet J. 2015;63: 9–16.
20. Whalin L, Weary DM, von Keyserlingk MAG. Understanding Behavioural Development of Calves in Natural Settings to Inform Calf Management. Animals (Basel). 2021;11. doi:10.3390/ani11082446
21. Schuck-Paim C, Alonso WJ, Hamilton C. Short agony or long ache: comparing sources of suffering that differ in duration and intensity. Effective Altruism Forum. 2024. Available: https://forum.effectivealtruism.org/editPost?postId=C2qiY9hwH3Xuirce3)
22. Berridge KC, Kringelbach ML. Neuroscience of affect: brain mechanisms of pleasure and displeasure. Curr Opin Neurobiol. 2013;23: 294–303.
23. Reimert I, Webb LE, van Marwijk MA, Bolhuis JE. Review: Towards an integrated concept of animal welfare. Animal. 2023;17 Suppl 4: 100838.
24. Poirier C, Bateson M, Gualtieri F, Armstrong EA, Laws GC, Boswell T, et al. Validation of hippocampal biomarkers of cumulative affective experience. Neurosci Biobehav Rev. 2019;101: 113–121.
25. Suffering and happiness: Morally symmetric or orthogonal? In: Center for Reducing Suffering [Internet]. 9 Sep 2020 [cited 29 Feb 2024]. Available: https://centerforreducingsuffering.org/research/suffering-and-happiness-morally-symmetric-or-orthogonal/
26. Shriver A. The Asymmetrical Contributions of Pleasure and Pain to Subjective Well-Being. RevPhilPsych. 2014;5: 135–153.
27. Tomasik B. Are Happiness and Suffering Symmetric? [cited 29 Feb 2024]. Available: https://reducing-suffering.org/happiness-suffering-symmetric/

Could Transparency International be a model for improved animal welfare?

editar — Sat, 24 Feb 2024 23:03:03 +0000

Could Transparency International Be a Model to Improve Farm Animal Welfare?

Cynthia Schuck-Paim, Wladimir J Alonso,

February 24, 2024
11:03 pm

In a recent article shared on the Effective Altruism Forum, we discuss a new way to help farm animals. Drawing inspiration from the work of Transparency International, we explore the potential of applying similar principles to improve farm animal welfare.

Key Points:

The Challenge of Transparency: Lack of transparency enables companies to circumvent reforms and prevents assessing the true effectiveness of animal welfare policies.
Promoting Accountability: Increased transparency would promote accountability, compliance with standards, and enhancement of welfare practices.
Proposed Initiatives: We propose the creation of a dedicated organization focused on enhancing transparency in animal welfare. An organization focused on transparency could develop reporting frameworks, auditing processes, traceability systems, sourcing policies requiring transparency, and welfare labeling schemes.
Engaging Consumers: Increased consumer awareness through transparent labeling, traceability, sourcing policies, and educational campaigns could help bridge the gap between preferences and realities.
Potential Areas of Work: Transparency initiatives could include public sharing of independent audit results, animal-based health monitoring, stockmanship qualifications, slaughter line inspections, and transparency rankings of companies.

Read the full article on the Effective Altruism Forum

Welfare Footprint Institute

Defining Affective Experiences

How to Define Affective Experiences for Analysis: Striking the Right Level of Detail

How to Decide What Level of Detail Is Appropriate

Some Useful Rules

Confounding Factors in Welfare Comparisons

Confounding Factors in Welfare Comparisons of Animal Production Systems

Transition to Animal Welfare Research

Resource and Management Investment Disparities

ANIMAL SELECTION BIASES

Early Life Experience and System Adaptation

Incomplete Welfare Outcome Coverage

Environmental and Geographic Variability

TEMPORAL CONSIDERATIONS

Examples of Confounding in Comparisons of Production Systems

laying hens

salmon aquaculture

beef cattle

PIGS

ANALYTICAL APPROACHES ADDRESSING COUNFOUNDING

CONCLUSION

Borrowing the Knowns and Unknowns Framework to Reveal Blind Spots in Animal Welfare Science

Borrowing the Knowns and Unknowns Framework to Reveal Blind Spots in Animal Welfare Science​

1. Known Knowns: What We Know We Know

2. Known Unknowns: What We Know We Don’t Know

3. Unknown Knowns: What We Don’t Realize We Already Know

4. Unknown Unknowns: What We Don’t Know We Don’t Know

An Important Way to Address Unknowns: Transparency About the Limitations of Results

Using the Welfare Footprint Framework to document animal welfare in legal settings

Using the Welfare Footprint Framework to Document Animal Welfare in Legal Settings

Our Scientific Director Discusses WFP’s Methodology on Hear This Idea

Quantifying Animal Welfare: Our Scientific Director Discusses WFP's Methodology on Hear This Idea

Bridging Neuroscience and Philosophy: Exploring Animal Emotions and Welfare with the Neurophilosopher GPT Tool

Bridging Neuroscience and Philosophy: Exploring Animal Emotions with the Neurophilosopher GPT Tool

Bridging Neuroscience and Philosophy: The Old Brain vs. New Brain

Introducing the Neurophilosopher GPT Tool

Practical Applications and Exploration

SOME RECOMMENDED READING SOURCES

Consulting Opportunity: Beef Cattle Welfare Research

Opportunity: Consulting Work on Beef Cattle Welfare

Key Responsibilities

who we are looking for

CONTRACT DETAILS

APPLICATION

Can AI power the Global Mapping and Quantification of Animal Suffering? The Pain Atlas Project

Can AI power the Global Mapping and Quantification of Animal Suffering? the Pain Atlas Project

Recruiting AI to help quantify animal suffering

The Pain ATLAS PROJECT

Quantifying Positive Animal Welfare in Individuals and Populations

Beyond Suffering: A Framework for Quantifying Positive Animal Welfare in Individuals and Populations

Main Takeaways

INTRODUCTION

Operational definition of pleasure

Categories of Pleasure Intensity

Pleasure-Tracks

Cumulative Pleasure (individual level)

Cumulative Pleasure (population level)

A notation proposal for Welfare Footprints

Further Thoughts

References

Could Transparency International be a model for improved animal welfare?

Could Transparency International Be a Model to Improve Farm Animal Welfare?

Borrowing the Knowns and Unknowns Framework to Reveal Blind Spots in Animal Welfare Science