# Scratch Pad Philosophy of Technology ## Do technologies merely instrumentally serve human ends, or do they help constitute moral agency?” by Rachele Baruzzo The aim of this essay is to address the following question: do technologies merely instrumentally serve human ends, or do they help constitute moral agency? I support the following thesis: technology helps constitute moral agency, **it can embody values but all responsibility lies with humans**, that are those involved in its design, implementation, and use; because only human beings have intentionality and phenomenological consciousness. The relationship, consequentially, is asymmetric and a distinction between agency (the power to act) and accountability (the duty to answer for that act) is fundamental. While technology shapes our actions and decisions, it can never be held responsible; morality resides within the relationship between technology and human actors (*relational normativity*). To provide a clearer understanding of this debate, I categorize the relevant philosophical perspectives into two groups: \-Proponents of a “distributed morality”-such as the philosophers Verbeek, Van den Hoven & Vermaas, and Parsons \-argue that morality is not an exclusively human attribute but is instead embedded and distributed across technologies. \-Proponents of a “human-centric thesis” \-such as the philosophers Noorman & Johnson- argue that moral responsibility remains closely tied to human intentionality and action. The essay is divided into two parts: first, I present the arguments in favor of my thesis, integrating the perspectives of major philosophers in the field; second I deepen the analysis by exploring the nuances of my argumentation and addressing the principal critiques \-such as the Neutrality thesis- leveled against the thesis I support. Finally I will present my conclusions. *1\. The Co-constitution of Moral Agency: Action Schemes and Affordances* Getting to the heart of the matter, the question is whether technologies are merely instruments, or, if they actively shape the way of seeing the world, the choices we make and the way of life. Technology mediates the human experience in two ways: perception and action. Regarding perception, the philosopher Verbeek(2014) in *“The Moral Status of Technical Artefacts”* exposes the example of obstetric ultrasound: by allowing parents to know about the child’s health prenatally, the technology provides a moral choice \-whether to continue the pregnancy or not- that before the invention of this technology was not even possible to take. Regarding action, technologies are not neutral: they “invite” or “inhibit” certain behaviors. Following Gibson’s(1979) concept of *affordances,* every object “invites” a specific type of action. For instance, holding a gun gives a different kind of “inclination” toward pulling a trigger than having nothing in one’s hand. While the human remains the one who pulls the trigger—as the gun lacks intentionality— yet it reconfigures the human’s moral environment: the presence or not of a gun alters the agent's possibilities of action. This thesis aligns closely with the concept of “action schemes” introduced by philosophers Illies & Meijers(2014). Within this framework, a technology is not responsible in itself; rather, it influences morality by “inviting” human agents toward specific actions. Although a hammer cannot “decide” to hit a nail, its physical structure promotes certain motions and discourages others. Moreover, the authors claim an artifact cannot be understood in isolation, but only within an “action scheme” that includes both technology and user. For instance, I argue, if a management software is designed to prioritize one gender over another; the action scheme it imposes carries a specific moral connotation that does not depend solely on the manager's individual will. In fact, responsibility is multifaceted and distributed across various domains: epistemic responsibility requires designers to understand how their creations works; managerial responsibility demands that those implementing technology remain aware of the values embedded in the software; and social responsibility implicates the entire network of stakeholders, including users who are aware of the system's biases, because they are responsible for accepting its use or failing to report it. Expanding on this argument, theorist and philosopher L. Winner(1980) maintains that artifacts can embody political and moral values. Through the analysis of Robert Moses’ low overpasses in New York, the author argues that an overpass designed low in order to prevent buses from reaching the beaches(since those who were using this transportation were usually African Americans) is a physical embodiment of an immoral decision. In fact, it perpetuates racial and social exclusion through a material form. Hence, morality resides in the relation between the technology and the entire network of actors—producers, sellers, and users alike. However, there are differences between humans and technology: humans possess consciousness, moral agency, intentionality, complex mental states…; attributes that technology entirely lacks. Technology is designed to help humans, and a “collateral effect” is that it inevitably embodies values, that is why we should be careful about what values we embed within them. Thus, morality lies in the relation between humans and technologies. While technology is essential to human life, because technology is also a sharp stone to hunt animals; the relationship remains asymmetric. We make it for our own purpose, *we* decide, therefore we are responsible. Who is *we*? The technology producer/designer who embeds the initial values, the seller that markets the tool, and the end-user who uses it in real life. Certainly someone is more responsible, someone less; this is \-in fact- distributed responsibility; we can refer to it also as a multi-layered structure of responsibility. Therefore, why do I argue technology can embody values? Everything humans use they can design and use it to convey positive or negative values, and even the object’s shape or color can change its use, or the values it carries. From the most basic level to the most advanced, humans must take responsibility for design and market. Let us consider epistemic bubbles and echo chambers, they shape our behavior in both our online and offline lives because they prevent us from seeing or understanding anything outside of them. Moreover, in social networks like Instagram, algorithms use Big Data to track all the users’ choices and clicks, showing them only what is relevant to their past interests. Consequently, users see only what the algorithm lets them see, encountering nothing that challenges their opinions. That is why, when we delegate a choice to a technology, we are not just using a tool: in this case, the algorithm utilizes a biased logic that constitutes a moral judgment. Moreover, for instance, if an algorithm denies a loan based on a person’s origin, it is performing a moral act of discrimination. Technology does help constitute moral agency, yet, who is responsible for it? We cannot point the finger at technology, because it has no intentionality and no free will: these two attributes are necessary for being held responsible. Technology is created by humans to serve humans. Therefore, again, humans are liable. I have argued technology constitutes how we live and choose, how it conveys values and shapes the world. About this argument the philosophers Van den Hoven & Vermaas(2015) maintain technology is value-laden, meaning that objects themselves embody ethical principles. The design of an object defines what is possible or impossible to do, influencing human behavior. The idea is that ethical dilemmas can be solved through better design solutions \-addressing issues before the technology is implemented, in order to avoid wasting time, money, and effort- rather than relying solely on rules or laws afterwards. This is why the authors introduce value sensitive design(VSD): technology is sensitive to the values humans embed within it. Consequently, our goal must be to design technology that promotes the common good. To achieve this, we must all bear responsibility and remember that we—not the machines—are in control. We should not blindly accept technology as "good" simply because it is popular or approved by influential figures. Even if an overpass is built with discriminatory intent(R. Moses), we can still choose to walk, to protest, or to redesign. Sometimes we prefer to pretend we are objects to avoid the anguish of our own freedom. We flee from choice, because freedom implies responsibility, which is a heavy burden. However, we must instead reclaim our role within the system and take charge of the technologies we design. To illustrate Value Sensitive Design further, consider another example. If an architect designs a new museum without providing easy access for people with disabilities \-such as ramps or elevators- that building embodies "bad" and exclusionary values. Once humans design an object, it begins to shape the world; this confirms that technology is inherently value-laden. In fact, moreover, if a consumer purchases a product from a brand that utilizes child labor, can that act be considered morally neutral? Does his choice to buy or not that product make a fundamental difference? Yet, if, on one hand, ethical choices are impossible without material mediation, and technology helps define our moral landscape; on the other hand we cannot hold technology itself accountable. Responsibility, I claim again, remains a human attribute. 2. *Addressing the Neutrality Thesis and Social Responsibility* According to the Neutrality thesis, technology is neutral and it’s seen as a value-free instrument that simply increases human efficiency. "Guns don't kill people, people kill people", is the most renowned thesis statement. However I argue that if technology were solely instrumental, the introduction of a new tool would not change the moral landscape; it would simply allow a pre-existing goal to be achieved more efficiently. Yet, we saw that technology creates new moral categories, for example with Verbeek’s(2014) example of obstetric ultrasound: it created a moral dilemma that could not exist before its creation. Therefore, the instrument did not just serve an end; it redefined the end itself. This means technology is not neutral. Let us consider another example given by Verbeek(2014): a guardrail on a road. It is not neutral, the philosopher claims: it *imposes* a safety choice and limits freedom of movement. By its very design, the guardrail limits the driver’s freedom of movement in favor of the moral value of preserving life. The philosophers Noorman & Johnson(2014) would argue, in contrast to Verbeek, that the "morality" is not in the guardrail itself, but entirely in the intentions of the engineers who designed it and the legal authority that installed it. For the authors, it is just a tool; the morality is one hundred per cent human. I argue, again, that morality is shared both in technology and humans, yet the responsibility remains exclusively human. Morality is shared because a guardrail or a speed bump(others examples from the author Verbeek(2014)) are in favour of a good value, and perpetuating it; they are actively shaping the world in which humans will choose. But if instead of the guardrail there is nothing or there is some malicious engineer that put one of his new designs of steel spikes on the road, this is perpetuating a disvalue. This complexity illustrates the intricate nature of morality, demonstrating that human ethical choices are inextricably linked to the technologies they bring into existence. To further dismantle the Neutrality Thesis, I align with Van de Poel & Kroes’ (2013) view that values are not just added later, but are embedded in the design process itself. The authors divide in intended value(value that the designer wants to be in technology), embodied value(value that the designer inscribed into technology) and realised value(value that is realised in the application and social integration of the technology). They argue that the value is embedded in technology, because the object has been intentionally designed to *promote* that value. For example, if the value is ‘freedom’ the designer might choose not to source components from countries that use those profits to fund wars or political imprisonment. I agree with this perspective, because it treats the embodiment of values as a relational property \-between the object, the designer's intentions, and the user's behaviour- rather than an intrinsic one(only inside the technology). Therefore, if technology embodies values, then design is not just an engineering challenge, if ever an ethical exercise \-as I already said. That is why, I argue, designers have a moral responsibility to anticipate how their technical choices will impact social values once the object is in the real world. Similarly, the philosopher Parsons(2015) claims that technology shapes the world: the design of fast cars creates the moral need for safety laws, just as digital technologies create privacy issues that previously didn't exist. Morality lies in the way a technology guides humans toward certain choices, the philosopher claims; though responsibility relies with people, as they are the only ones endowed with intentionality. Therefore, for the author, morality is found in the relation: a design can make certain moral actions easier or harder (e.g., inclusive architecture, vs. exclusive one). In this instance, I want to mention de Ruiter’s paper, ‘*The Distinct wrong of Deepfakes*’(2021). His thesis, briefly, is that Deepfake is not intrinsically evil, yet it is how we humans decide to use it that leads to immorality. Deepfake can also be used for good: it performs a better interaction with the students in the school, in the museums; it provides a voice to people with ALS, it protects the identity of political activists, and so on. Thus, deepfake is not evil in itself, but in the way humans use it. Therefore, it is important not to look only at the regulation after the release of technology, but at first at the design/project regulation. Every human being has the responsibility to carefully examine their actions and educate himself to promote positive values, because technology does help constitute moral agency. Expanding on this topic, the philosopher B. Preston(2013) claims responsibility lies within the social system. The author argues that technology structures reality, because it contributes to moral agency by organizing entire “systems of action”, not simply by providing predetermined purposes. Preston claims, as other authors such as Brey(2014), Floridi(2013), and \-as we saw Verbeek(2014), morality is distributed. Every choice human beings make comes entirely from social practices. A low overpass \-coming back to L. Winner analysis of Moses’ overpass in New York- is just stones put together, until a social system uses it to enforce segregation. Therefore, responsibility lies within the collective rules and social practices that are repeated over time. Morality emerges, Preston argues, when we use the technology in accordance with (or in violation of) the expectations that the community has established for that object. If an object is morally questionable, the "fault" lies not only with the person who designed it, but with the social system that continues to reproduce it and integrate it into practices. Morality, in short, emerges from these interconnected relationships. *Conclusions* I have argued that technology helps constitute moral agency, yet responsibility lies solely with humans, as they have intentionality and freedom of choice. Humans create technology to serve their own purposes and welfare. Throughout this discussion, I have introduced the views of philosophers such as Illies & Meijers, Noorman & Johnson, Van den Hoven & Vermaas, Van de Poel & Kroes, Parsons, Preston and Verbeek while providing counterarguments to the Neutrality Thesis. I have supported a perspective of hybrid morality: morality resides not only in technology, and not only in humans, but in the interaction between them; and not only in humans, yet in all the connections and relationships humans have, that is to say then, in the social system. Humans make decisions through and with technology. As argued, technology defines the scope of our possible actions. We do not just use it; we inhabit the world through it. Morality is not in the technology only, but in the relationship mediated by it, and technology does not only enable us to do new things, yet it imposes new dilemmas and responsibilities, becoming an integral part of our moral fabric, such as we saw in Verbeek's obstetric ultrasound example. Nevertheless, we must be careful not to treat technology as an autonomous agent or suggest that morality is equally distributed between humans and technology; because if we do, we downright lose the responsibility, and the evident difference between who made the technology and the technology itself. If we treat technology as a co-equal moral agent, right at the same level as humans; we forget the asymmetric relation and we risk forgetting our power to change technology; we might excuse its harms by claiming we are not fully accountable. Technology constitutes a large portion of our world, but it never “chooses” *for* us. Responsibility only relies on us. However, it seems to me that humans are trying all the fairytale ways to flee from responsibility… References Brey, P. (2014). “From Moral Agents to Moral Factors: The Structural Ethics Approach”, in *The Moral Status of Technical Artefacts*, (pp. 125–142). Springer. De Ruiter, A. (2021). “The Distinct Wrong of Deepfakes”, *Philosophy & Technology,* 34(4), 1311–1332. Floridi, L. (2013). “Distributed Morality in an Information Society”, *Science and Engineering Ethics* 19(3), 727–743. Gibson, J.J. (1979). “The Ecological Approach to Visual Perception”, Houghton Mifflin. Johnson, D. G. & Noorman, M. (2014). “Artefactual Agency and Artefactual Moral Agency”, in *The Moral Status of Technical Artefacts*, Springer. Houkes, W. & Vermaas, P. E. (2010). “Technical Functions: On the Use and Design of Artefacts”, chapter 2 (“*Use, Plans and Designing*”). Illies, C. F. R. & Meijers, A. (2014). “Artefacts, Agency, and Action Schemes”, in *The Moral Status of Technical Artefacts*, Springer. Parsons, G. (2015). “The Philosophy of Design”, chapter 7 (“*Ethics*”). Preston, B. (2013). “A Philosophy of Material Culture: Action, Function, and Mind”. Chapter 5 (*“Proper Function and System Function*”). Van den Hoven, J., Vermaas, P. E. & Van de Poel, I. (2015). “Handbook of Ethics and Values in Technological Design”. Springer Van de Poel, I., & Kroes, P. (2014). “Can technology embody values?” In P. Kroes & P.-P. Verbeek (Eds.), *The Moral Status of Technical Artefacts* (pp. 103–124). Springer Netherlands. Verbeek, P.-P. (2014). “Some Misunderstandings About the Moral Significance of Technology”, in P. Kroes & P.-P. Verbeek (eds.), *The Moral Status of Technical Artefacts*, Springer. Winner, L. (1980). “Do Artifacts Have Politics?"**,** *Daedalus* (Vol. 109, No. 1, pp. 121–136). --- # What's Happening *Active threads and today's activity — updated by /harvest* ## Active ## Sessions ## Actions --- ## Environmental Aesthetics of AI - provisional section structure Provisional section-facing version, not a settled draft. Sections 2 and 3 have left us with a trained continuation system whose generated text is not made aesthetically intelligible by either design knowledge or person-directed knowledge. Design knowledge explains the conditions under which the system is made and trained. Person-directed knowledge explains why the output can seem like the speech of an interlocutor. Neither gives the right kind of knowledge for appreciating the generated continuation as the product of a trained continuation system. Carlson's general demand still applies. Appropriate appreciation depends on knowledge suited to the kind of thing the object is. In the natural case, order appreciation is guided by knowledge of the processes that produce visible order. A cliff face is not appreciated well by treating it as if a designer had arranged its layers. Its visible form becomes intelligible when it is seen as the result of geological formation. The knowledge matters because it changes what the visible form can be seen as. The analogous knowledge for LLM outputs cannot be knowledge of inner mechanism alone. Mechanistic interpretability may tell us something important about how a model works, but its objects are not the generated text as encountered by a reader. Circuits, attention heads, and activation patterns are not what the appreciator reads. They may explain the system at one level, but they do not by themselves make the visible order of the continuation aesthetically available. The needed knowledge must operate at the level of generated linguistic form. It must explain how a trained system carries text forward from a context, and how the resulting continuation comes to have the shape it has. The relevant object is the response as a formed continuation: text that develops from a prompt under learned and imposed constraints. Semiotic physics is the name I will use for this kind of knowledge. It is an account of how generated text is carried forward by a trained continuation system. It concerns the order of linguistic forms as they are taken up, continued, transformed, and constrained over the course of generation. This order is path-dependent. A prompt gives the model an initial context, but the continuation also conditions itself as it develops. Later parts of the response are shaped by earlier parts of the response. A generated text can gather a direction, maintain it, or lose it because each step changes the context from which the next step is produced. This also explains why the appearances discussed in Section 3 remain aesthetically salient. A generated continuation may seem patient, evasive, playful, or severe. Those appearances need not be treated as traits of a subject. They can be understood as profiles sustained in text: patterns in how the continuation carries forward a posture, a tempo, or a relation to the prompt. The language of forces is useful only in this restricted sense. It does not attribute agency to the model. It marks regularities, constraints, and dependence on prior state. A force, here, is a factor that helps explain why the continuation goes on as it does. Knowing that a text is LLM-generated therefore changes how it is read. The same sentence can be understood as part of an authored utterance, or as part of a continuation shaped by context, learned regularities, and post-training. Semiotic physics matters aesthetically when it supports the second mode of aspection. It lets the generated text appear as the visible order of a trained continuation system. That order can be encountered at more than one scale. A single output is a bounded continuation from a context. A chat is an extended exchange in which earlier turns shape later ones. A model is the trained system whose tendencies become visible across outputs and chats. These are different levels at which semiotic order can become appreciable. ## Enrico feedback on Environmental Aesthetics of AI Today Hello Nick, I read the old version and re-read the new one, do you have time for a short zoom or wa chat today? 10:26 Otherwise we'll discuss all this tomorrow 10:26 I start writing here my comments. 10:36 I agree that the overarching structure of the new one is much better but I still think that the sections on artifacts and persons were better in the previous version. Whe should give more empahsis to Olah's quotations from the beginning. The scaffold-light-growing triad is the key. 10:39 The scaffold cass for design appreciation, the "light" somehow calls for a sort of person appreciation, but what is missing between the two is "growing" (training, emergence) that rather calls for order appreciation. 10:41 Here is how I see the paper now: 1) Begin with Olah's quotation, stressing that there are three key elements of LLMs: scaffold, "light" and growing. 2) Move to Carlson's design appreciation, showing that it only accounts for scaffold. 3) Introduce person appreciation to (at least partially) account fo "light", with qualificaiton (the "assistant role"). 4) Argue that the combination of design appreciation (as for the scaffold) and person appreciation (as for the "light") is still missing something crucial: growing. Here is where we need order appreciation. 5) Conclude that the aesthetics of LLM mobilizes not just one dimension (design/person/order) but all three. 10:48 La metafora più famosa sull'lA non è più valida - THE BUNKER Sono passati cinque anni dalla pubblicazione di On the Dangers of Stochastic Parrots, e Emily Bender è tornata sulla formula che aveva coniato con un post intitolato Stochastic Parrots: Frequently Unasked Questions. Il testo si presenta come chiarimento filologico, una sorta di servizio reso al pubblico che, nel frattempo, ha trasformato l'espressione in slogan di [...] ® the-bunker.it 11:07 This passage from our old versions sounds great to me, I think this should be the key to the whole paper: "'we create the scaffold that it grows on and we create the light that it grows towards. But the thing that we actually create, it's this almost biological entity or organism that we're studying'. (Olah 2024) Carlson's recommendation to appreciate things for what they in fact are might seem to warn against taking such a comparison seriously - LLMs are not biological organisms, and their 'growth' is a computational process of parameter adjustment, not biological development. But the comparison is apt: the organisation of a trained neural network is not specified by its designers but emerges from a process they set in motion. This distinguishes LLMs from traditional designed artifacts, and, as we shall see, it has consequences for what kind of appreciation is appropriate." 11:16 I wonder whether we should look for a specific scientific discipline (semiotic physics vs mechaniastic interpetability) or just try to characterize the relevant knowledge required to carry out order appreciation for LLMs regardless of the label and the disciplinary framing. 11:28 The key for me is this passage of the old (PPR-submitted) version: Despite their different framings - Janus's simulator ontology and Picca's Peircean semiotics - these accounts share a core insight: we should attend to what regularities govern how text propagates through the system, not to whether LLMs think or intend. In a similar vein, Wolfram (2023) states that inside ChatGPT any piece of text is effectively represented by an array of numbers that we can think of as coordinates of a point in some kind of 'linguistic feature space. So when ChatGPT continues a piece of text this corresponds to tracing out a trajectory in linguistic feature space. But now we can ask what makes this trajectory correspond to text we consider meaningful. And might there perhaps be some kind of 'semantic laws of motion! From Wolfram's perspective, semiotic physics thus would have three main objects to investigate: (i) the "linguistic feature space" in which words and other linguistic items have their place; (ii) the "trajectories" that can be traced out in this space to continue a piece of text; and (iii) the "semantic laws of motion" that determine such trajectories. We draw on this literature but develop it in a specific direction. Our aim is to show how semiotic physics can serve as the "right kind of knowledge" for aesthetic appreciation of LLMs in Carlson's sense: the knowledge that makes order visible and intelligible, and that guides acts of aspection. The connection to environmental aesthetics, and the claim that semiotic physics can play the role for LLMs that geology plays for landscapes, is our contribution. We also articulate the 'forces' of semiotic physics at the level of textual effects rather than at the level of mechanistic detail. The existing literature tends to discuss semiotic physics in terms of probability distributions, embedding spaces, and dynamical systems. These descriptions are accurate, but they do not directly connect to what readers can perceive in generated text. Our articulation of the forces operates at a level that does connect to perceivable features. What does semiotic physics track? The regularities it describes manifest as perceivable features of generated text. Consider vocabulary clustering: words do not appear independently but make other related words more probable, so that once a medical term appears, other medical terms become more likely to follow. Or consider coherence dynamics: the model threads mate... Read more 11:31 Perhaps we don't even need to call all this "semiotic physics", we can just say that this is the relevant knowledge for order appreciation of LLMs 11:32 This passage of our old version also seems very important: "The aspection guided by semiotic physics is, in a sense, aspection of language itself - of the textual order produced by semiotic forces. We are not attending to mechanical internals - activation patterns, attention weights, circuit-level features - since these require specialist tools and are not accessible to readers. We are attending to the textual manifestation of semiotic order: how vocabulary clusters, how coherence is maintained or lost across an exchange, how register persists or shifts, how post-training shapes response structure. These are features of the language itself, perceivable by competent readers. Competent readers already have tacit knowledge of how language works: syntactic, semantic, pragmatic, and discourse-level knowledge built up through immersion in spoken and written language. They perceive patterns in LLM outputs using this tacit knowledge. Semiotic physics adds explicit articulation of these patterns and a causal story about their source in training. The competent reader senses that different models have different 'vibes'; semiotic physics-knowledge of how meaning clusters, how register persists, how training shapes the texture of response-explains what produces those vibes and makes them available for sustained attention." 11:38 The whole section 5.2 (Practical Acquaintance) also is very important because it makes order appreication of LLMs intelligible, it avoids that the relevant sort of knowledge (semiotic physics") remained mysterious. 11:41 (Do you know the Italian expression "Buttare via il bambino con l'acqua sporca"?) 11:43 This passage (end of 6.3) also seems to say something very important that we should not downplay and probably discuss earleier, when talking about the relevant knowldege for order appreciaiton of LLMs: "Each model, as an instantiation of semiotic physics, has learned regularities from human text. It reflects, in transformed form, the semiotic culture of its training data. There is a sense in which generative Al is a mirror of culture, not only morally, as Vallor (2024) has argued, but aesthetically. The model shows us our own semiotic patterns, filtered through statistical learning. Appreciating an LLM is, in part, appreciating culture seen through technology." 11:52 To wrap up, the set-up should something along this lines: We have the aesthetics of design (Carlson and Parsons, Forsey), let's try to apply it to LLMs. It doesn't work because of Olah's point. Let's try first to supplement it with person aesthetics. It's a bit better, but not satisfying yet, it still misses something crucial. We need environmantal aesthetics to miss the gap 1g erial. We nee environmatal aesthetics to mister ga's between design (captued by design aesthetics) and function (somwhow partlally captured by person aesthetics). And we need to higlight the relevant knowledge for order appreciation of such environments. 12:00 *fill the gap 12:01 "But the thing that we actually create, it's this almost biological entity or organism that we're studying'. (Olah 2024)" We shouls stress that "almost biological entity" might be interpreted in two ways: person and environent. The person-interpretation only capures something about the function (with due qualifications), whereas the environment-interpretation capture the core features that fill the gap between design and personish functioning. 12:12