Pub Date : 2022-02-17DOI: 10.1093/oso/9780197552384.003.0003
Oron Shagrir
This chapter focuses on Turing’s analysis, which reduces effective computability to Turing machine computability. The analysis consists of two steps. The first is the formulation of restrictive conditions on effective computations by idealized human computers. The second is an outline of a reduction theorem, from human computability to Turing machine computability. The chapter discusses in detail the claim that the Turing analysis targets human computers, and explicates some aspects of this notion. It then argues that while this analysis of human computability is of immense theoretical and practical importance, it cannot be taken as the basis of machine computation.
{"title":"Turing’s Computability","authors":"Oron Shagrir","doi":"10.1093/oso/9780197552384.003.0003","DOIUrl":"https://doi.org/10.1093/oso/9780197552384.003.0003","url":null,"abstract":"This chapter focuses on Turing’s analysis, which reduces effective computability to Turing machine computability. The analysis consists of two steps. The first is the formulation of restrictive conditions on effective computations by idealized human computers. The second is an outline of a reduction theorem, from human computability to Turing machine computability. The chapter discusses in detail the claim that the Turing analysis targets human computers, and explicates some aspects of this notion. It then argues that while this analysis of human computability is of immense theoretical and practical importance, it cannot be taken as the basis of machine computation.","PeriodicalId":222167,"journal":{"name":"The Nature of Physical Computation","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124305141","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-02-17DOI: 10.1093/oso/9780197552384.003.0009
Oron Shagrir
The chapter advances an argument for the semantic individuation of computational states. Its first and central premise is the simultaneous implementation of automata by physical systems. Simultaneous implementation (also known as “the indeterminacy of computation”) is a phenomenon whereby a physical system implements multiple formal structures at the same time, at the same location, and even with the same physical properties. The next premises are that, in a given context, a computational taxonomy often takes into account one of the implemented formalisms, and that content determines, at least partly, the select formalism. The chapter then addresses two objections. One is that the computational structure of the system is always identified with a more basic structure. The second objection is that we need not appeal to content for the purposes of computational individuation because extrinsic yet non-semantic features would do the job just as well.
{"title":"An Argument for the Semantic View","authors":"Oron Shagrir","doi":"10.1093/oso/9780197552384.003.0009","DOIUrl":"https://doi.org/10.1093/oso/9780197552384.003.0009","url":null,"abstract":"The chapter advances an argument for the semantic individuation of computational states. Its first and central premise is the simultaneous implementation of automata by physical systems. Simultaneous implementation (also known as “the indeterminacy of computation”) is a phenomenon whereby a physical system implements multiple formal structures at the same time, at the same location, and even with the same physical properties. The next premises are that, in a given context, a computational taxonomy often takes into account one of the implemented formalisms, and that content determines, at least partly, the select formalism. The chapter then addresses two objections. One is that the computational structure of the system is always identified with a more basic structure. The second objection is that we need not appeal to content for the purposes of computational individuation because extrinsic yet non-semantic features would do the job just as well.","PeriodicalId":222167,"journal":{"name":"The Nature of Physical Computation","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125210926","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-02-17DOI: 10.1093/oso/9780197552384.003.0006
Oron Shagrir
Many accounts of computation associate it with the implementation of some abstract structure such as an automaton, algorithm, or program. This chapter focuses on David Chalmers’s account. Chalmers develops an account of implementation as a response to the so-called triviality results, which indicate that every physical object performs every computation, and he goes on to identify computation with implementation. The chapter presents a nuanced approach to Chalmers’s account. On the positive side, it suggests that his notion of implementation successfully circumvents the dire consequences of the triviality results. Additionally, it agrees that the implementation of some formalism is necessary for computing. On the negative side, it argues that implementation (in Chalmers’s sense) is not sufficient for computation. Many physical systems—such as rocks, stomachs, and hurricanes—do not compute, even when they implement a formalism of some sort.
{"title":"Computation as Implementation","authors":"Oron Shagrir","doi":"10.1093/oso/9780197552384.003.0006","DOIUrl":"https://doi.org/10.1093/oso/9780197552384.003.0006","url":null,"abstract":"Many accounts of computation associate it with the implementation of some abstract structure such as an automaton, algorithm, or program. This chapter focuses on David Chalmers’s account. Chalmers develops an account of implementation as a response to the so-called triviality results, which indicate that every physical object performs every computation, and he goes on to identify computation with implementation. The chapter presents a nuanced approach to Chalmers’s account. On the positive side, it suggests that his notion of implementation successfully circumvents the dire consequences of the triviality results. Additionally, it agrees that the implementation of some formalism is necessary for computing. On the negative side, it argues that implementation (in Chalmers’s sense) is not sufficient for computation. Many physical systems—such as rocks, stomachs, and hurricanes—do not compute, even when they implement a formalism of some sort.","PeriodicalId":222167,"journal":{"name":"The Nature of Physical Computation","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123670194","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-02-17DOI: 10.1093/oso/9780197552384.003.0004
Oron Shagrir
This chapter distinguishes between three kinds of machine computation, and characterizes the inclusion relations between them. The most general notion is that of generic computation. A second one is that of algorithmic computation, and the third, which is the focus of the book, is that of physical computation. The chapter starts with Robin Gandy’s characterization of machine computation, and argues that the account falls ambiguously between the different kinds of machine computation and fails to fully capture any of them. Next, the chapter analyzes the notions of generic, algorithmic, and physical computation, concluding that these notions are extensionally distinct.
{"title":"Preamble to Machine Computation","authors":"Oron Shagrir","doi":"10.1093/oso/9780197552384.003.0004","DOIUrl":"https://doi.org/10.1093/oso/9780197552384.003.0004","url":null,"abstract":"This chapter distinguishes between three kinds of machine computation, and characterizes the inclusion relations between them. The most general notion is that of generic computation. A second one is that of algorithmic computation, and the third, which is the focus of the book, is that of physical computation. The chapter starts with Robin Gandy’s characterization of machine computation, and argues that the account falls ambiguously between the different kinds of machine computation and fails to fully capture any of them. Next, the chapter analyzes the notions of generic, algorithmic, and physical computation, concluding that these notions are extensionally distinct.","PeriodicalId":222167,"journal":{"name":"The Nature of Physical Computation","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124610209","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-02-17DOI: 10.1093/oso/9780197552384.003.0002
Oron Shagrir
The chapter outlines the various demands that arise in a philosophical account of physical computation. It discusses two lists of desiderata put forward by Smith (1996, 2002) and Piccinini (2007, 2015), and argues that a key desideratum is to formulate classification criteria that distinguish computing from non-computing physical systems. It further suggests that the distinction between these criteria for distinguishing computing from non-computing systems (the-right-things-compute and the-wrong-things-don’t-compute) might differ from the criteria for classifying different kinds of computation. It also argues that we need not require substantive demands about the objectivity and the utility of computation in advance.
{"title":"Desiderata of a Theory of Computation","authors":"Oron Shagrir","doi":"10.1093/oso/9780197552384.003.0002","DOIUrl":"https://doi.org/10.1093/oso/9780197552384.003.0002","url":null,"abstract":"The chapter outlines the various demands that arise in a philosophical account of physical computation. It discusses two lists of desiderata put forward by Smith (1996, 2002) and Piccinini (2007, 2015), and argues that a key desideratum is to formulate classification criteria that distinguish computing from non-computing physical systems. It further suggests that the distinction between these criteria for distinguishing computing from non-computing systems (the-right-things-compute and the-wrong-things-don’t-compute) might differ from the criteria for classifying different kinds of computation. It also argues that we need not require substantive demands about the objectivity and the utility of computation in advance.","PeriodicalId":222167,"journal":{"name":"The Nature of Physical Computation","volume":"128 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133461657","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-02-17DOI: 10.1093/oso/9780197552384.003.0008
Oron Shagrir
The semantic view of computation asserts that semantic properties are an essential aspect of the nature of physical computing systems. This chapter starts with an explication of the semantic view of computation: its claims, its variants, how it differs from its non-semantic counterparts, what semantics means in computational contexts, and the major arguments in its favor. Next, the chapter addresses some of the main objections to the semantic view, advanced by Frances Egan, Gualtiero Piccinini, Michael Rescorla, and others. It argues that when more refined distinctions are made, the semantic view can overcome these objections.
{"title":"The Semantic View of Computation","authors":"Oron Shagrir","doi":"10.1093/oso/9780197552384.003.0008","DOIUrl":"https://doi.org/10.1093/oso/9780197552384.003.0008","url":null,"abstract":"The semantic view of computation asserts that semantic properties are an essential aspect of the nature of physical computing systems. This chapter starts with an explication of the semantic view of computation: its claims, its variants, how it differs from its non-semantic counterparts, what semantics means in computational contexts, and the major arguments in its favor. Next, the chapter addresses some of the main objections to the semantic view, advanced by Frances Egan, Gualtiero Piccinini, Michael Rescorla, and others. It argues that when more refined distinctions are made, the semantic view can overcome these objections.","PeriodicalId":222167,"journal":{"name":"The Nature of Physical Computation","volume":"32 16","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114114537","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-02-17DOI: 10.1093/oso/9780197552384.003.0007
Oron Shagrir
The mechanistic account has evolved into a formidable theory of physical computation, and is the dominant view of computation today. The chapter focuses on Gualtiero Piccinini’s account, which is the most comprehensive and detailed theory of physical computation to date. According to this theory, computation is a functional mechanism. Its teleological function is to manipulate vehicles based solely on differences between different portions of the vehicles according to a rule defined over the vehicles. After presenting the account, the chapter highlights its two main shortcomings. One is that computational explanations do not sit squarely with the mechanistic framework. The other is that the main criteria of the account—rules, medium-independence, and teleological functions—do not appear to adequately distinguish computing from non-computing systems. The conclusion is that, despite its obvious virtues, the mechanistic account falls short of satisfying the key classification and explanation desiderata of an account of computation.
{"title":"Computation as Mechanism","authors":"Oron Shagrir","doi":"10.1093/oso/9780197552384.003.0007","DOIUrl":"https://doi.org/10.1093/oso/9780197552384.003.0007","url":null,"abstract":"The mechanistic account has evolved into a formidable theory of physical computation, and is the dominant view of computation today. The chapter focuses on Gualtiero Piccinini’s account, which is the most comprehensive and detailed theory of physical computation to date. According to this theory, computation is a functional mechanism. Its teleological function is to manipulate vehicles based solely on differences between different portions of the vehicles according to a rule defined over the vehicles. After presenting the account, the chapter highlights its two main shortcomings. One is that computational explanations do not sit squarely with the mechanistic framework. The other is that the main criteria of the account—rules, medium-independence, and teleological functions—do not appear to adequately distinguish computing from non-computing systems. The conclusion is that, despite its obvious virtues, the mechanistic account falls short of satisfying the key classification and explanation desiderata of an account of computation.","PeriodicalId":222167,"journal":{"name":"The Nature of Physical Computation","volume":"142 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128185137","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-02-17DOI: 10.1093/oso/9780197552384.003.0010
Oron Shagrir
This chapter argues that input-output modeling is an essential element of computing, at least in current computational approaches in cognitive neuroscience. A model, in the context of this work, is a representational system that preserves patterns of relations of the target domain. A process input-output models a given target when its input-output function and some relation in the target have a shared formal structure. Finally, the chapter shows that modeling is often associated with computing, that it plays a major methodological role in discovering what function is being computed, and that it enhances a distinctive account of computational explanation.
{"title":"Computing as Modeling","authors":"Oron Shagrir","doi":"10.1093/oso/9780197552384.003.0010","DOIUrl":"https://doi.org/10.1093/oso/9780197552384.003.0010","url":null,"abstract":"This chapter argues that input-output modeling is an essential element of computing, at least in current computational approaches in cognitive neuroscience. A model, in the context of this work, is a representational system that preserves patterns of relations of the target domain. A process input-output models a given target when its input-output function and some relation in the target have a shared formal structure. Finally, the chapter shows that modeling is often associated with computing, that it plays a major methodological role in discovering what function is being computed, and that it enhances a distinctive account of computational explanation.","PeriodicalId":222167,"journal":{"name":"The Nature of Physical Computation","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122427456","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-02-17DOI: 10.1093/oso/9780197552384.003.0005
Oron Shagrir
The chapter deals with Robert Cummins’s account of computation. In this account, to compute is to execute a program, and program execution is reduced to step-satisfaction. The main claim of this chapter is that step-satisfaction is not a necessary feature of computation. The more general moral is that select architectural profiles, such as step-satisfaction, do not help to distinguish between computing and non-computing. Depending on how it is understood, step-satisfaction either excludes important cases of computing physical systems or is empty, applying to virtually every physical system. The argument rests on the analysis of two examples. One is a thermal device for averaging numbers. The other is an attractor neural network that solves the n-queens problem.
{"title":"Computation as Step-Satisfaction","authors":"Oron Shagrir","doi":"10.1093/oso/9780197552384.003.0005","DOIUrl":"https://doi.org/10.1093/oso/9780197552384.003.0005","url":null,"abstract":"The chapter deals with Robert Cummins’s account of computation. In this account, to compute is to execute a program, and program execution is reduced to step-satisfaction. The main claim of this chapter is that step-satisfaction is not a necessary feature of computation. The more general moral is that select architectural profiles, such as step-satisfaction, do not help to distinguish between computing and non-computing. Depending on how it is understood, step-satisfaction either excludes important cases of computing physical systems or is empty, applying to virtually every physical system. The argument rests on the analysis of two examples. One is a thermal device for averaging numbers. The other is an attractor neural network that solves the n-queens problem.","PeriodicalId":222167,"journal":{"name":"The Nature of Physical Computation","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127942673","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-02-17DOI: 10.1093/oso/9780197552384.003.0011
Oron Shagrir
The chapter recaps the proposed characterization of physical computation and outlines how this characterization squares with the desiderata of an account of computation. The account meets the desideratum of the-right-things-compute in that it deems smartphones, laptops, and robots, as well as natural cognitive and nervous systems, to be computing systems. It meets the desideratum of the-wrong-things-don’t-compute in that it deems stomachs, hurricanes, and rocks, for instance, to be non-computing systems. The proposed account meets a milder objectivity desideratum. It is consistent with the claims that the computational properties of some computing systems, such as brains, are entirely objective and that some computational properties of all computing systems are entirely objective (PO2). The account assigns a distinct feature to computational explanations, namely, modeling, and it emphasizes that the identity conditions that define types of computation (whose specification is the goal of the taxonomy desideratum) are different from those conditions that distinguish computation from non-computation. Finally, it is said how the account fares with Smith’s scope criteria.
{"title":"Conclusion","authors":"Oron Shagrir","doi":"10.1093/oso/9780197552384.003.0011","DOIUrl":"https://doi.org/10.1093/oso/9780197552384.003.0011","url":null,"abstract":"The chapter recaps the proposed characterization of physical computation and outlines how this characterization squares with the desiderata of an account of computation. The account meets the desideratum of the-right-things-compute in that it deems smartphones, laptops, and robots, as well as natural cognitive and nervous systems, to be computing systems. It meets the desideratum of the-wrong-things-don’t-compute in that it deems stomachs, hurricanes, and rocks, for instance, to be non-computing systems. The proposed account meets a milder objectivity desideratum. It is consistent with the claims that the computational properties of some computing systems, such as brains, are entirely objective and that some computational properties of all computing systems are entirely objective (PO2). The account assigns a distinct feature to computational explanations, namely, modeling, and it emphasizes that the identity conditions that define types of computation (whose specification is the goal of the taxonomy desideratum) are different from those conditions that distinguish computation from non-computation. Finally, it is said how the account fares with Smith’s scope criteria.","PeriodicalId":222167,"journal":{"name":"The Nature of Physical Computation","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117066473","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：