Toward Human-Level Artificial Intelligence Representation and Computation of Meaning in Natural Language

by Jackson, Philip C., Jr.

• ISBN13: 9780486833002
• ISBN10: 0486833003
• Format: Paperback
• Copyright: 2019-11-13
• Publisher: Dover Publications

Summary

How can human-level artificial intelligence be achieved? What are the potential consequences? This book describes a research approach toward achieving human-level AI, combining a doctoral thesis and research papers by the author.
The research approach, called TalaMind, involves developing an AI system that uses a 'natural language of thought' based on the unconstrained syntax of a language such as English; designing the system as a collection of concepts that can create and modify concepts to behave intelligently in an environment; and using methods from cognitive linguistics for multiple levels of mental representation. Proposing a design-inspection alternative to the Turing Test, these pages discuss 'higher-level mentalities' of human intelligence, which include natural language understanding, higher-level forms of learning and reasoning, imagination, and consciousness. Dr. Jackson gives a comprehensive review of other research, addresses theoretical objections to the proposed approach and to achieving human-level AI in principle, and describes a prototype system that illustrates the potential of the approach.
This book discusses economic risks and benefits of AI, considers how to ensure that human-level AI and superintelligence will be beneficial for humanity, and gives reasons why human-level AI may be necessary for humanity's survival and prosperity.

Author Biography

Philip C. Jackson, Jr. has a Ph.D. for a thesis on artificial intelligence, and has extensive experience in software design and development. He is the founder of TalaMind LLC, conducting research toward human-level artificial intelligence.

Table of Contents

Figures. xii
§ Notation and Overview of Changes. xiii
Synopsis. xv
Preface. xvii
1.        Introduction. 1
1.1      Can Machines Have Human-Level Intelligence?. 1
1.2      Thesis Approach. 4
1.3      Terminology: Tala and TalaMind. 7
1.4      TalaMind Hypotheses. 7
1.4.1       Intelligence Kernel Hypothesis. 8
1.4.2       Natural Language Mentalese Hypothesis. 9
1.4.3       Multiple Levels of Mentality Hypothesis. 11
1.4.4       Relation to the Physical Symbol System Hypothesis. 11
1.5      TalaMind System Architecture. 12
1.6      Arguments and Evidence: Strategy and Criteria for Success. 16
1.7      Overview of Chapters. 18
2.        Subject Review: Human-Level AI and Natural Language. 19
2.1      Human-Level Artificial Intelligence. 19
2.1.1       How to Define and Recognize Human-Level AI 19
2.1.2       Unexplained Features of Human-Level Intelligence. 22
2.1.2.1     Generality. 22
2.1.2.2     Creativity and Originality. 23
2.1.2.3     Natural Language Understanding. 24
2.1.2.4     Effectiveness, Robustness, Efficiency. 24
2.1.2.5     Self-Development and Higher-Level Learning. 24
2.1.2.6     Metacognition and Multi-Level Reasoning. 25
2.1.2.7     Imagination. 26
2.1.2.8     Consciousness. 26
2.1.2.9     Sociality, Emotions, Values. 27
2.1.2.10     Visualization, Spatial-Temporal Reasoning. 28
2.1.2.11     Curiosity, Self-Programming, Theory of Mind. 28
Figures. viii
§ Notation and Overview of Changes. ix
Synopsis. xi
Preface. xiii
1.        Introduction. 1
1.1      Can Machines Have Human-Level Intelligence?. 1
1.2      Thesis Approach. 4
1.3      Terminology: Tala and TalaMind. 7
1.4      TalaMind Hypotheses. 7
1.4.1       Intelligence Kernel Hypothesis. 8
1.4.2       Natural Language Mentalese Hypothesis. 9
1.4.3       Multiple Levels of Mentality Hypothesis. 11
1.4.4       Relation to the Physical Symbol System Hypothesis. 11
1.5      TalaMind System Architecture. 12
1.6      Arguments and Evidence: Strategy and Criteria for Success. 16
1.7      Overview of Chapters. 18
2.        Subject Review: Human-Level AI and Natural Language. 19
2.1      Human-Level Artificial Intelligence. 19
2.1.1       How to Define and Recognize Human-Level AI 19
2.1.2       Unexplained Features of Human-Level Intelligence. 22
2.1.2.1     Generality. 22
2.1.2.2     Creativity and Originality. 23
2.1.2.3     Natural Language Understanding. 24
2.1.2.4     Effectiveness, Robustness, Efficiency. 24
2.1.2.5     Self-Development and Higher-Level Learning. 24
2.1.2.6     Metacognition and Multi-Level Reasoning. 25
2.1.2.7     Imagination. 26
2.1.2.8     Consciousness. 26
2.1.2.9     Sociality, Emotions, Values. 27
2.1.2.10     Visualization, Spatial-Temporal Reasoning. 28
2.1.2.11     Curiosity, Self-Programming, Theory of Mind. 28
2.1.2.12     Other Unexplained Features. 29
2.2      Natural Language. 29
2.2.1       Does Thought Require Language?. 29
2.2.2       What Does Meaning Mean?. 33
2.2.3       Does Human-Level AI Require Embodiment?. 37
2.2.4       Natural Language, Metacognition, Inner Speech. 39
2.3      Relation of Thesis Approach to Previous Research. 41
2.3.1       Formal, Logical Approaches. 41
2.3.2       Cognitive Approaches and Cognitive Linguistics. 42
2.3.3       Approaches to Human-Level Artificial Intelligence. 45
2.3.3.1     Sloman. 45
2.3.3.2     Minsky. 45
2.3.3.3     McCarthy. 47
2.3.3.4     Reverse-Engineering the Brain. 48
2.3.3.5     Cognitive Architectures and AGI 49
2.3.3.6     Newell and Simon’s Cognitive Research. 50
2.3.3.6.1     Unified Theories of Cognition. 50
2.3.3.6.2     The ‘Knowledge Level’ and ‘Intelligence Level’ 51
2.3.3.7     Other Influences for Thesis Approach. 53
2.3.4       Approaches to Artificial Consciousness. 53
2.3.5       Approaches to Reflection and Self-Programming. 55
2.3.6       Johnson-Laird’s Mental Models. 59
2.3.7       Research on Natural Logic. 61
2.3.7.1     Natural Logic According to Lakoff 61
2.3.7.2     Monotonicity-Based Natural Logic. 62
2.4      Summary. 64
3.        Analysis of Thesis Approach to Human-Level AI 65
3.1      Overview.. 65
3.2      Theoretical Requirements for TalaMind Architecture. 66
3.2.1       Conceptual Language. 66
3.2.2       Conceptual Framework. 70
3.2.3       Conceptual Processes. 72
3.3      Representing Meaning with Natural Language Syntax. 73
3.4      Representing English Syntax in Tala. 76
3.4.1       Non-Prescriptive, Open, Flexible. 76
3.4.2       Semantic and Ontological Neutrality and Generality. 77
3.5      Choices and Methods for Representing English Syntax. 77
3.5.1       Theoretical Approach to Represent English Syntax. 77
Figures. viii
§ Notation and Overview of Changes. ix
Synopsis. xi
Preface. xiii
1.        Introduction. 1
1.1      Can Machines Have Human-Level Intelligence?. 1
1.2      Thesis Approach. 4
1.3      Terminology: Tala and TalaMind. 7
1.4      TalaMind Hypotheses. 7
1.4.1       Intelligence Kernel Hypothesis. 8
1.4.2       Natural Language Mentalese Hypothesis. 9
1.4.3       Multiple Levels of Mentality Hypothesis. 11
1.4.4       Relation to the Physical Symbol System Hypothesis. 11
1.5      TalaMind System Architecture. 12
1.6      Arguments and Evidence: Strategy and Criteria for Success. 16
1.7      Overview of Chapters. 18
2.        Subject Review: Human-Level AI and Natural Language. 19
2.1      Human-Level Artificial Intelligence. 19
2.1.1       How to Define and Recognize Human-Level AI 19
2.1.2       Unexplained Features of Human-Level Intelligence. 22
2.1.2.1     Generality. 22
2.1.2.2     Creativity and Originality. 23
2.1.2.3     Natural Language Understanding. 24
2.1.2.4     Effectiveness, Robustness, Efficiency. 24
2.1.2.5     Self-Development and Higher-Level Learning. 24
2.1.2.6     Metacognition and Multi-Level Reasoning. 25
2.1.2.7     Imagination. 26
2.1.2.8     Consciousness. 26
2.1.2.9     Sociality, Emotions, Values. 27
2.1.2.10     Visualization, Spatial-Temporal Reasoning. 28
2.1.2.11     Curiosity, Self-Programming, Theory of Mind. 28
2.1.2.12     Other Unexplained Features. 29
2.2      Natural Language. 29
2.2.1       Does Thought Require Language?. 29
2.2.2       What Does Meaning Mean?. 33
2.2.3       Does Human-Level AI Require Embodiment?. 37
2.2.4       Natural Language, Metacognition, Inner Speech. 39
2.3      Relation of Thesis Approach to Previous Research. 41
2.3.1       Formal, Logical Approaches. 41
2.3.2       Cognitive Approaches and Cognitive Linguistics. 42
2.3.3       Approaches to Human-Level Artificial Intelligence. 45
2.3.3.1     Sloman. 45
2.3.3.2     Minsky. 45
2.3.3.3     McCarthy. 47
2.3.3.4     Reverse-Engineering the Brain. 48
2.3.3.5     Cognitive Architectures and AGI 49
2.3.3.6     Newell and Simon’s Cognitive Research. 50
2.3.3.6.1     Unified Theories of Cognition. 50
2.3.3.6.2     The ‘Knowledge Level’ and ‘Intelligence Level’ 51
2.3.3.7     Other Influences for Thesis Approach. 53
2.3.4       Approaches to Artificial Consciousness. 53
2.3.5       Approaches to Reflection and Self-Programming. 55
2.3.6       Johnson-Laird’s Mental Models. 59
2.3.7       Research on Natural Logic. 61
2.3.7.1     Natural Logic According to Lakoff 61
2.3.7.2     Monotonicity-Based Natural Logic. 62
2.4      Summary. 64
3.        Analysis of Thesis Approach to Human-Level AI 65
3.1      Overview.. 65
3.2      Theoretical Requirements for TalaMind Architecture. 66
3.2.1       Conceptual Language. 66
3.2.2       Conceptual Framework. 70
3.2.3       Conceptual Processes. 72
3.3      Representing Meaning with Natural Language Syntax. 73
3.4      Representing English Syntax in Tala. 76
3.4.1       Non-Prescriptive, Open, Flexible. 76
3.4.2       Semantic and Ontological Neutrality and Generality. 77
3.5      Choices and Methods for Representing English Syntax. 77
3.5.1       Theoretical Approach to Represent English Syntax. 77
3.5.2       Representing Syntactic Structure of NL Sentences. 78
3.6      Semantic Representation and Processing. 80
3.6.1       Lexemes, Senses, Referents, and Variables. 80
3.6.2       Multiple Representations for the Same Concept 83
3.6.3       Representing Interpretations. 84
3.6.3.1     Underspecification. 85
3.6.3.2     Syntactic Elimination of Interpretations. 85
3.6.3.3     Generic and Non-Generic Interpretations. 87
3.6.3.4     Specific and Non-Specific Interpretations. 87
3.6.3.5     Individual and Collective Interpretations. 87
3.6.3.6     Count and Mass Interpretations. 87
3.6.3.7     Quantificational Interpretations. 88
3.6.3.8     De Dicto and De Re Interpretations. 90
3.6.3.9     Interpretations of Compound Noun Structures. 92
3.6.3.10     Interpretations of Metaphors. 93
3.6.3.11     Interpretations of Metonyms. 93
3.6.3.12     Interpretations of Anaphora. 93
3.6.3.13     Interpretation of Idioms. 94
3.6.4       Semantic Disambiguation. 95
3.6.5       Representing Implications. 95
3.6.6       Semantic Inference. 96
3.6.6.1     Representation of Truth. 97
3.6.6.2     Negation and Contradictions. 97
3.6.6.3     Inference with Commonsense. 101
3.6.6.4     Paraphrase and Inference. 102
3.6.6.5     Inference for Metaphors and Metonyms. 102
3.6.7       Representation of Contexts. 103
3.6.7.1     Dimensions of Context 103
3.6.7.2     Perceived Reality. 107
3.6.7.3     Event Memory. 107
3.6.7.4     Encyclopedic and Commonsense Knowledge. 107
3.6.7.5     Interactive Contexts and Mutual Knowledge. 109
3.6.7.6     Hypothetical Contexts. 112
3.6.7.7     Semantic Domains. 113
3.6.7.8     Mental Spaces. 114
3.6.7.9     Conceptual Blends. 118
3.6.7.10     Theory Contexts. 121
3.6.7.11     Problem Contexts. 123
3.6.7.12     Composite Contexts. 124
3.6.7.13     Society of Mind Thought Context 124
3.6.7.14     Meta-Contexts. 124
3.6.8 Primitive Words and Variables in Tala  125
Figures. viii
§ Notation and Overview of Changes. ix
Synopsis. xi
Preface. xiii
1.        Introduction. 1
1.1      Can Machines Have Human-Level Intelligence?. 1
1.2      Thesis Approach. 4
1.3      Terminology: Tala and TalaMind. 7
1.4      TalaMind Hypotheses. 7
1.4.1       Intelligence Kernel Hypothesis. 8
1.4.2       Natural Language Mentalese Hypothesis. 9
1.4.3       Multiple Levels of Mentality Hypothesis. 11
1.4.4       Relation to the Physical Symbol System Hypothesis. 11
1.5      TalaMind System Architecture. 12
1.6      Arguments and Evidence: Strategy and Criteria for Success. 16
1.7      Overview of Chapters. 18
2.        Subject Review: Human-Level AI and Natural Language. 19
2.1      Human-Level Artificial Intelligence. 19
2.1.1       How to Define and Recognize Human-Level AI 19
2.1.2       Unexplained Features of Human-Level Intelligence. 22
2.1.2.1     Generality. 22
2.1.2.2     Creativity and Originality. 23
2.1.2.3     Natural Language Understanding. 24
2.1.2.4     Effectiveness, Robustness, Efficiency. 24
2.1.2.5     Self-Development and Higher-Level Learning. 24
2.1.2.6     Metacognition and Multi-Level Reasoning. 25
2.1.2.7     Imagination. 26
2.1.2.8     Consciousness. 26
2.1.2.9     Sociality, Emotions, Values. 27
2.1.2.10     Visualization, Spatial-Temporal Reasoning. 28
2.1.2.11     Curiosity, Self-Programming, Theory of Mind. 28
2.1.2.12     Other Unexplained Features. 29
2.2      Natural Language. 29
2.2.1       Does Thought Require Language?. 29
2.2.2       What Does Meaning Mean?. 33
2.2.3       Does Human-Level AI Require Embodiment?. 37
2.2.4       Natural Language, Metacognition, Inner Speech. 39
2.3      Relation of Thesis Approach to Previous Research. 41
2.3.1       Formal, Logical Approaches. 41
2.3.2       Cognitive Approaches and Cognitive Linguistics. 42
2.3.3       Approaches to Human-Level Artificial Intelligence. 45
2.3.3.1     Sloman. 45
2.3.3.2     Minsky. 45
2.3.3.3     McCarthy. 47
2.3.3.4     Reverse-Engineering the Brain. 48
2.3.3.5     Cognitive Architectures and AGI 49
2.3.3.6     Newell and Simon’s Cognitive Research. 50
2.3.3.6.1     Unified Theories of Cognition. 50
2.3.3.6.2     The ‘Knowledge Level’ and ‘Intelligence Level’ 51
2.3.3.7     Other Influences for Thesis Approach. 53
2.3.4       Approaches to Artificial Consciousness. 53
2.3.5       Approaches to Reflection and Self-Programming. 55
2.3.6       Johnson-Laird’s Mental Models. 59
2.3.7       Research on Natural Logic. 61
2.3.7.1     Natural Logic According to Lakoff 61
2.3.7.2     Monotonicity-Based Natural Logic. 62
2.4      Summary. 64
3.        Analysis of Thesis Approach to Human-Level AI 65
3.1      Overview.. 65
3.2      Theoretical Requirements for TalaMind Architecture. 66
3.2.1       Conceptual Language. 66
3.2.2       Conceptual Framework. 70
3.2.3       Conceptual Processes. 72
3.3      Representing Meaning with Natural Language Syntax. 73
3.4      Representing English Syntax in Tala. 76
3.4.1       Non-Prescriptive, Open, Flexible. 76
3.4.2       Semantic and Ontological Neutrality and Generality. 77
3.5      Choices and Methods for Representing English Syntax. 77
3.5.1       Theoretical Approach to Represent English Syntax. 77
3.5.2       Representing Syntactic Structure of NL Sentences. 78
3.6      Semantic Representation and Processing. 80
3.6.1       Lexemes, Senses, Referents, and Variables. 80
3.6.2       Multiple Representations for the Same Concept 83
3.6.3       Representing Interpretations. 84
3.6.3.1     Underspecification. 85
3.6.3.2     Syntactic Elimination of Interpretations. 85
3.6.3.3     Generic and Non-Generic Interpretations. 87
3.6.3.4     Specific and Non-Specific Interpretations. 87
3.6.3.5     Individual and Collective Interpretations. 87
3.6.3.6     Count and Mass Interpretations. 87
3.6.3.7     Quantificational Interpretations. 88
3.6.3.8     De Dicto and De Re Interpretations. 90
3.6.3.9     Interpretations of Compound Noun Structures. 92
3.6.3.10     Interpretations of Metaphors. 93
3.6.3.11     Interpretations of Metonyms. 93
3.6.3.12     Interpretations of Anaphora. 93
3.6.3.13     Interpretation of Idioms. 94
3.6.4       Semantic Disambiguation. 95
3.6.5       Representing Implications. 95
3.6.6       Semantic Inference. 96
3.6.6.1     Representation of Truth. 97
3.6.6.2     Negation and Contradictions. 97
3.6.6.3     Inference with Commonsense. 101
3.6.6.4     Paraphrase and Inference. 102
3.6.6.5     Inference for Metaphors and Metonyms. 102
3.6.7       Representation of Contexts. 103
3.6.7.1     Dimensions of Context 103
3.6.7.2     Perceived Reality. 107
3.6.7.3     Event Memory. 107
3.6.7.4     Encyclopedic and Commonsense Knowledge. 107
3.6.7.5     Interactive Contexts and Mutual Knowledge. 109
3.6.7.6     Hypothetical Contexts. 112
3.6.7.7     Semantic Domains. 113
3.6.7.8     Mental Spaces. 114
3.6.7.9     Conceptual Blends. 118
3.6.7.10     Theory Contexts. 121
3.6.7.11     Problem Contexts. 123
3.6.7.12     Composite Contexts. 124
3.6.7.13     Society of Mind Thought Context 124
3.6.7.14     Meta-Contexts. 124
3.6.8       Primitive Words and Variables in Tala. 125
3.7      Higher-Level Mentalities. 128
3.7.1       Multi-Level Reasoning. 129
3.7.1.1     Deduction. 129
3.7.1.2     Induction. 129
3.7.1.3     Abduction. 129
3.7.1.4     Analogical Reasoning. 130
3.7.1.5     Causal and Purposive Reasoning. 130
3.7.1.6     Meta-Reasoning. 131
3.7.2       Self-Development and Higher-Level Learning. 131
3.7.2.1     Learning by Multi-Level Reasoning. 131
3.7.2.2     Learning by Reflection and Self-Programming. 132
3.7.2.3     Learning by Invention of Languages. 134
3.7.3       Curiosity. 135
3.7.4       Imagination. 138
3.7.5       Sociality, Emotions, Values. 139
3.7.6       Consciousness. 139
3.8      Summary. 141
4.        Theoretical Issues and Objections. 142
4.1      Issues and Objections re the Possibility of Human-Level AI 142
4.1.1       Dreyfus Issues. 142
4.1.2       Penrose Objections. 144
4.1.2.1     General Claims re Intelligence. 144
4.1.2.2     Claims re Human Logical Insight 146
4.1.2.3     Gödelian Arguments. 147
4.1.2.4     Continuous Computation. 154
4.1.2.5     Hypothesis re Orchestrated Objective Reduction. 154
4.2      Issues and Objections for Thesis Approach. 156
4.2.1       Theoretical Objections to a Language of Thought 156
4.2.2       Objections to Representing Semantics via NL Syntax. 156
4.2.2.1     The Circularity Objection. 156
4.2.2.2     Objection Syntax Is Insufficient for Semantics. 156
4.2.2.3     Ambiguity Objections to Natural Language. 157
4.2.2.4     Objection Thought Is Perceptual, Not Linguistic. 158
4.2.3       Weizenbaum’s Eliza Program.. 159
4.2.4       Searle’s Chinese Room Argument 161
4.2.5       McCarthy’s Objections to Natural Language Mentalese. 164
4.2.6       Minsky’s Issues for Representation and Learning. 167
4.2.7       Chalmers’ Hard Problem of Consciousness. 169
4.2.8       Smith’s Issues for Representation and Reflection. 172
4.3      Summary. 177
Figures. viii
§ Notation and Overview of Changes. ix
Synopsis. xi
Preface. xiii
1.        Introduction. 1
1.1      Can Machines Have Human-Level Intelligence?. 1
1.2      Thesis Approach. 4
1.3      Terminology: Tala and TalaMind. 7
1.4      TalaMind Hypotheses. 7
1.4.1       Intelligence Kernel Hypothesis. 8
1.4.2       Natural Language Mentalese Hypothesis. 9
1.4.3       Multiple Levels of Mentality Hypothesis. 11
1.4.4       Relation to the Physical Symbol System Hypothesis. 11
1.5      TalaMind System Architecture. 12
1.6      Arguments and Evidence: Strategy and Criteria for Success. 16
1.7      Overview of Chapters. 18
2.        Subject Review: Human-Level AI and Natural Language. 19
2.1      Human-Level Artificial Intelligence. 19
2.1.1       How to Define and Recognize Human-Level AI 19
2.1.2       Unexplained Features of Human-Level Intelligence. 22
2.1.2.1     Generality. 22
2.1.2.2     Creativity and Originality. 23
2.1.2.3     Natural Language Understanding. 24
2.1.2.4     Effectiveness, Robustness, Efficiency. 24
2.1.2.5     Self-Development and Higher-Level Learning. 24
2.1.2.6     Metacognition and Multi-Level Reasoning. 25
2.1.2.7     Imagination. 26
2.1.2.8     Consciousness. 26
2.1.2.9     Sociality, Emotions, Values. 27
2.1.2.10     Visualization, Spatial-Temporal Reasoning. 28
2.1.2.11     Curiosity, Self-Programming, Theory of Mind. 28
2.1.2.12     Other Unexplained Features. 29
2.2      Natural Language. 29
2.2.1       Does Thought Require Language?. 29
2.2.2       What Does Meaning Mean?. 33
2.2.3       Does Human-Level AI Require Embodiment?. 37
2.2.4       Natural Language, Metacognition, Inner Speech. 39
2.3      Relation of Thesis Approach to Previous Research. 41
2.3.1       Formal, Logical Approaches. 41
2.3.2       Cognitive Approaches and Cognitive Linguistics. 42
2.3.3       Approaches to Human-Level Artificial Intelligence. 45
2.3.3.1     Sloman. 45
2.3.3.2     Minsky. 45
2.3.3.3     McCarthy. 47
2.3.3.4     Reverse-Engineering the Brain. 48
2.3.3.5     Cognitive Architectures and AGI 49
2.3.3.6     Newell and Simon’s Cognitive Research. 50
2.3.3.6.1     Unified Theories of Cognition. 50
2.3.3.6.2     The ‘Knowledge Level’ and ‘Intelligence Level’ 51
2.3.3.7     Other Influences for Thesis Approach. 53
2.3.4       Approaches to Artificial Consciousness. 53
2.3.5       Approaches to Reflection and Self-Programming. 55
2.3.6       Johnson-Laird’s Mental Models. 59
2.3.7       Research on Natural Logic. 61
2.3.7.1     Natural Logic According to Lakoff 61
2.3.7.2     Monotonicity-Based Natural Logic. 62
2.4      Summary. 64
3.        Analysis of Thesis Approach to Human-Level AI 65
3.1      Overview.. 65
3.2      Theoretical Requirements for TalaMind Architecture. 66
3.2.1       Conceptual Language. 66
3.2.2       Conceptual Framework. 70
3.2.3       Conceptual Processes. 72
3.3      Representing Meaning with Natural Language Syntax. 73
3.4      Representing English Syntax in Tala. 76
3.4.1       Non-Prescriptive, Open, Flexible. 76
3.4.2       Semantic and Ontological Neutrality and Generality. 77
3.5      Choices and Methods for Representing English Syntax. 77
3.5.1       Theoretical Approach to Represent English Syntax. 77
3.5.2       Representing Syntactic Structure of NL Sentences. 78
3.6      Semantic Representation and Processing. 80
3.6.1       Lexemes, Senses, Referents, and Variables. 80
3.6.2       Multiple Representations for the Same Concept 83
3.6.3       Representing Interpretations. 84
3.6.3.1     Underspecification. 85
3.6.3.2     Syntactic Elimination of Interpretations. 85
3.6.3.3     Generic and Non-Generic Interpretations. 87
3.6.3.4     Specific and Non-Specific Interpretations. 87
3.6.3.5     Individual and Collective Interpretations. 87
3.6.3.6     Count and Mass Interpretations. 87
3.6.3.7     Quantificational Interpretations. 88
3.6.3.8     De Dicto and De Re Interpretations. 90
3.6.3.9     Interpretations of Compound Noun Structures. 92
3.6.3.10     Interpretations of Metaphors. 93
3.6.3.11     Interpretations of Metonyms. 93
3.6.3.12     Interpretations of Anaphora. 93
3.6.3.13     Interpretation of Idioms. 94
3.6.4       Semantic Disambiguation. 95
3.6.5       Representing Implications. 95
3.6.6       Semantic Inference. 96
3.6.6.1     Representation of Truth. 97
3.6.6.2     Negation and Contradictions. 97
3.6.6.3     Inference with Commonsense. 101
3.6.6.4     Paraphrase and Inference. 102
3.6.6.5     Inference for Metaphors and Metonyms. 102
3.6.7       Representation of Contexts. 103
3.6.7.1     Dimensions of Context 103
3.6.7.2     Perceived Reality. 107
3.6.7.3     Event Memory. 107
3.6.7.4     Encyclopedic and Commonsense Knowledge. 107
3.6.7.5     Interactive Contexts and Mutual Knowledge. 109
3.6.7.6     Hypothetical Contexts. 112
3.6.7.7     Semantic Domains. 113
3.6.7.8     Mental Spaces. 114
3.6.7.9     Conceptual Blends. 118
3.6.7.10     Theory Contexts. 121
3.6.7.11     Problem Contexts. 123
3.6.7.12     Composite Contexts. 124
3.6.7.13     Society of Mind Thought Context 124
3.6.7.14     Meta-Contexts. 124
3.6.8       Primitive Words and Variables in Tala. 125
3.7      Higher-Level Mentalities. 128
3.7.1       Multi-Level Reasoning. 129
3.7.1.1     Deduction. 129
3.7.1.2     Induction. 129
3.7.1.3     Abduction. 129
3.7.1.4     Analogical Reasoning. 130
3.7.1.5     Causal and Purposive Reasoning. 130
3.7.1.6     Meta-Reasoning. 131
3.7.2       Self-Development and Higher-Level Learning. 131
3.7.2.1     Learning by Multi-Level Reasoning. 131
3.7.2.2     Learning by Reflection and Self-Programming. 132
3.7.2.3     Learning by Invention of Languages. 134
3.7.3       Curiosity. 135
3.7.4       Imagination. 138
3.7.5       Sociality, Emotions, Values. 139
3.7.6       Consciousness. 139
3.8      Summary. 141
4.        Theoretical Issues and Objections. 142
4.1      Issues and Objections re the Possibility of Human-Level AI 142
4.1.1       Dreyfus Issues. 142
4.1.2       Penrose Objections. 144
4.1.2.1     General Claims re Intelligence. 144
4.1.2.2     Claims re Human Logical Insight 146
4.1.2.3     Gödelian Arguments. 147
4.1.2.4     Continuous Computation. 154
4.1.2.5     Hypothesis re Orchestrated Objective Reduction. 154
4.2      Issues and Objections for Thesis Approach. 156
4.2.1       Theoretical Objections to a Language of Thought 156
4.2.2       Objections to Representing Semantics via NL Syntax. 156
4.2.2.1     The Circularity Objection. 156
4.2.2.2     Objection Syntax Is Insufficient for Semantics. 156
4.2.2.3     Ambiguity Objections to Natural Language. 157
4.2.2.4     Objection Thought Is Perceptual, Not Linguistic. 158
4.2.3       Weizenbaum’s Eliza Program.. 159
4.2.4       Searle’s Chinese Room Argument 161
4.2.5       McCarthy’s Objections to Natural Language Mentalese. 164
4.2.6       Minsky’s Issues for Representation and Learning. 167
4.2.7       Chalmers’ Hard Problem of Consciousness. 169
4.2.8       Smith’s Issues for Representation and Reflection. 172
4.3      Summary. 177
5.        Design of a Demonstration System... 178
5.1      Overview.. 178
5.2      Nature of the Demonstration System.. 179
5.3      Design of Conceptual Language. 180
5.3.1       Tala Syntax Notation. 181
5.3.2       Nouns. 182
5.3.3       Verbs. 185
5.3.4       Prepositions. 188
5.3.5       Pronouns. 190
5.3.6       Determiners. 190
5.3.7       Adjectives. 192
5.3.8       Adverbs. 193
5.3.9       Conjunctions. 193
5.3.9.1     Coordinating Conjunctions. 194
5.3.9.2     Subordinating / Structured Conjunctions. 196
5.3.9.3     Correlative Conjunctions. 198
5.3.10         Interjections. 199
5.3.11         Tala Variables and Pointers. 199
5.3.12         Inflections. 199
5.3.12.1     Determiner-Complement Agreement 200
5.3.12.2     Subject-Verb Agreement 200
5.4      Design of Conceptual Framework. 201
5.4.1       Requirements for a Conceptual Framework. 201
5.4.2       Structure of the Conceptual Framework. 202
5.4.3       Perceived Reality – Percepts and Effepts. 204
5.4.4       Subagents, Mpercepts, and Meffepts. 205
5.4.5       Tala Lexicon. 205
5.4.6       Encyclopedic Knowledge and Semantic Domains. 206
5.4.7       Current Domains. 207
5.4.8       Mental Spaces and Conceptual Blends. 207
5.4.9       Scenarios. 207
5.4.10         Thoughts. 208
5.4.11         Goals. 208
5.4.12         Executable Concepts. 208
5.4.13         Tala Constructions and Metaphors. 209
5.4.14         Event-Memory. 209
5.4.15         Systems. 209
5.4.16         The Reserved Variable ?self.. 209
5.4.17         Virtual Environment 210
5.5      Design of Conceptual Processes. 211
5.5.1       TalaMind Control Flow.. 211
Figures. viii
§ Notation and Overview of Changes. ix
Synopsis. xi
Preface. xiii
1.        Introduction. 1
1.1      Can Machines Have Human-Level Intelligence?. 1
1.2      Thesis Approach. 4
1.3      Terminology: Tala and TalaMind. 7
1.4      TalaMind Hypotheses. 7
1.4.1       Intelligence Kernel Hypothesis. 8
1.4.2       Natural Language Mentalese Hypothesis. 9
1.4.3       Multiple Levels of Mentality Hypothesis. 11
1.4.4       Relation to the Physical Symbol System Hypothesis. 11
1.5      TalaMind System Architecture. 12
1.6      Arguments and Evidence: Strategy and Criteria for Success. 16
1.7      Overview of Chapters. 18
2.        Subject Review: Human-Level AI and Natural Language. 19
2.1      Human-Level Artificial Intelligence. 19
2.1.1       How to Define and Recognize Human-Level AI 19
2.1.2       Unexplained Features of Human-Level Intelligence. 22
2.1.2.1     Generality. 22
2.1.2.2     Creativity and Originality. 23
2.1.2.3     Natural Language Understanding. 24
2.1.2.4     Effectiveness, Robustness, Efficiency. 24
2.1.2.5     Self-Development and Higher-Level Learning. 24
2.1.2.6     Metacognition and Multi-Level Reasoning. 25
2.1.2.7     Imagination. 26
2.1.2.8     Consciousness. 26
2.1.2.9     Sociality, Emotions, Values. 27
2.1.2.10     Visualization, Spatial-Temporal Reasoning. 28
2.1.2.11     Curiosity, Self-Programming, Theory of Mind. 28
2.1.2.12     Other Unexplained Features. 29
2.2      Natural Language. 29
2.2.1       Does Thought Require Language?. 29
2.2.2       What Does Meaning Mean?. 33
2.2.3       Does Human-Level AI Require Embodiment?. 37
2.2.4       Natural Language, Metacognition, Inner Speech. 39
2.3      Relation of Thesis Approach to Previous Research. 41
2.3.1       Formal, Logical Approaches. 41
2.3.2       Cognitive Approaches and Cognitive Linguistics. 42
2.3.3       Approaches to Human-Level Artificial Intelligence. 45
2.3.3.1     Sloman. 45
2.3.3.2     Minsky. 45
2.3.3.3     McCarthy. 47
2.3.3.4     Reverse-Engineering the Brain. 48
2.3.3.5     Cognitive Architectures and AGI 49
2.3.3.6     Newell and Simon’s Cognitive Research. 50
2.3.3.6.1     Unified Theories of Cognition. 50
2.3.3.6.2     The ‘Knowledge Level’ and ‘Intelligence Level’ 51
2.3.3.7     Other Influences for Thesis Approach. 53
2.3.4       Approaches to Artificial Consciousness. 53
2.3.5       Approaches to Reflection and Self-Programming. 55
2.3.6       Johnson-Laird’s Mental Models. 59
2.3.7       Research on Natural Logic. 61
2.3.7.1     Natural Logic According to Lakoff 61
2.3.7.2     Monotonicity-Based Natural Logic. 62
2.4      Summary. 64
3.        Analysis of Thesis Approach to Human-Level AI 65
3.1      Overview.. 65
3.2      Theoretical Requirements for TalaMind Architecture. 66
3.2.1       Conceptual Language. 66
3.2.2       Conceptual Framework. 70
3.2.3       Conceptual Processes. 72
3.3      Representing Meaning with Natural Language Syntax. 73
3.4      Representing English Syntax in Tala. 76
3.4.1       Non-Prescriptive, Open, Flexible. 76
3.4.2       Semantic and Ontological Neutrality and Generality. 77
3.5      Choices and Methods for Representing English Syntax. 77
3.5.1       Theoretical Approach to Represent English Syntax. 77
3.5.2       Representing Syntactic Structure of NL Sentences. 78
3.6      Semantic Representation and Processing. 80
3.6.1       Lexemes, Senses, Referents, and Variables. 80
3.6.2       Multiple Representations for the Same Concept 83
3.6.3       Representing Interpretations. 84
3.6.3.1     Underspecification. 85
3.6.3.2     Syntactic Elimination of Interpretations. 85
3.6.3.3     Generic and Non-Generic Interpretations. 87
3.6.3.4     Specific and Non-Specific Interpretations. 87
3.6.3.5     Individual and Collective Interpretations. 87
3.6.3.6     Count and Mass Interpretations. 87
3.6.3.7     Quantificational Interpretations. 88
3.6.3.8     De Dicto and De Re Interpretations. 90
3.6.3.9     Interpretations of Compound Noun Structures. 92
3.6.3.10     Interpretations of Metaphors. 93
3.6.3.11     Interpretations of Metonyms. 93
3.6.3.12     Interpretations of Anaphora. 93
3.6.3.13     Interpretation of Idioms. 94
3.6.4       Semantic Disambiguation. 95
3.6.5       Representing Implications. 95
3.6.6       Semantic Inference. 96
3.6.6.1     Representation of Truth. 97
3.6.6.2     Negation and Contradictions. 97
3.6.6.3     Inference with Commonsense. 101
3.6.6.4     Paraphrase and Inference. 102
3.6.6.5     Inference for Metaphors and Metonyms. 102
3.6.7       Representation of Contexts. 103
3.6.7.1     Dimensions of Context 103
3.6.7.2     Perceived Reality. 107
3.6.7.3     Event Memory. 107
3.6.7.4     Encyclopedic and Commonsense Knowledge. 107
3.6.7.5     Interactive Contexts and Mutual Knowledge. 109
3.6.7.6     Hypothetical Contexts. 112
3.6.7.7     Semantic Domains. 113
3.6.7.8     Mental Spaces. 114
3.6.7.9     Conceptual Blends. 118
3.6.7.10     Theory Contexts. 121
3.6.7.11     Problem Contexts. 123
3.6.7.12     Composite Contexts. 124
3.6.7.13     Society of Mind Thought Context 124
3.6.7.14     Meta-Contexts. 124
3.6.8       Primitive Words and Variables in Tala. 125
3.7      Higher-Level Mentalities. 128
3.7.1       Multi-Level Reasoning. 129
3.7.1.1     Deduction. 129
3.7.1.2     Induction. 129
3.7.1.3     Abduction. 129
3.7.1.4     Analogical Reasoning. 130
3.7.1.5     Causal and Purposive Reasoning. 130
3.7.1.6     Meta-Reasoning. 131
3.7.2       Self-Development and Higher-Level Learning. 131
3.7.2.1     Learning by Multi-Level Reasoning. 131
3.7.2.2     Learning by Reflection and Self-Programming. 132
3.7.2.3     Learning by Invention of Languages. 134
3.7.3       Curiosity. 135
3.7.4       Imagination. 138
3.7.5       Sociality, Emotions, Values. 139
3.7.6       Consciousness. 139
3.8      Summary. 141
4.        Theoretical Issues and Objections. 142
4.1      Issues and Objections re the Possibility of Human-Level AI 142
4.1.1       Dreyfus Issues. 142
4.1.2       Penrose Objections. 144
4.1.2.1     General Claims re Intelligence. 144
4.1.2.2     Claims re Human Logical Insight 146
4.1.2.3     Gödelian Arguments. 147
4.1.2.4     Continuous Computation. 154
4.1.2.5     Hypothesis re Orchestrated Objective Reduction. 154
4.2      Issues and Objections for Thesis Approach. 156
4.2.1       Theoretical Objections to a Language of Thought 156
4.2.2       Objections to Representing Semantics via NL Syntax. 156
4.2.2.1     The Circularity Objection. 156
4.2.2.2     Objection Syntax Is Insufficient for Semantics. 156
4.2.2.3     Ambiguity Objections to Natural Language. 157
4.2.2.4     Objection Thought Is Perceptual, Not Linguistic. 158
4.2.3       Weizenbaum’s Eliza Program.. 159
4.2.4       Searle’s Chinese Room Argument 161
4.2.5       McCarthy’s Objections to Natural Language Mentalese. 164
4.2.6       Minsky’s Issues for Representation and Learning. 167
4.2.7       Chalmers’ Hard Problem of Consciousness. 169
4.2.8       Smith’s Issues for Representation and Reflection. 172
4.3      Summary. 177
5.        Design of a Demonstration System... 178
5.1      Overview.. 178
5.2      Nature of the Demonstration System.. 179
5.3      Design of Conceptual Language. 180
5.3.1       Tala Syntax Notation. 181
5.3.2       Nouns. 182
5.3.3       Verbs. 185
5.3.4       Prepositions. 188
5.3.5       Pronouns. 190
5.3.6       Determiners. 190
5.3.7       Adjectives. 192
5.3.8       Adverbs. 193
5.3.9       Conjunctions. 193
5.3.9.1     Coordinating Conjunctions. 194
5.3.9.2     Subordinating / Structured Conjunctions. 196
5.3.9.3     Correlative Conjunctions. 198
5.3.10         Interjections. 199
5.3.11         Tala Variables and Pointers. 199
5.3.12         Inflections. 199
5.3.12.1     Determiner-Complement Agreement 200
5.3.12.2     Subject-Verb Agreement 200
5.4      Design of Conceptual Framework. 201
5.4.1       Requirements for a Conceptual Framework. 201
5.4.2       Structure of the Conceptual Framework. 202
5.4.3       Perceived Reality – Percepts and Effepts. 204
5.4.4       Subagents, Mpercepts, and Meffepts. 205
5.4.5       Tala Lexicon. 205
5.4.6       Encyclopedic Knowledge and Semantic Domains. 206
5.4.7       Current Domains. 207
5.4.8       Mental Spaces and Conceptual Blends. 207
5.4.9       Scenarios. 207
5.4.10         Thoughts. 208
5.4.11         Goals. 208
5.4.12         Executable Concepts. 208
5.4.13         Tala Constructions and Metaphors. 209
5.4.14         Event-Memory. 209
5.4.15         Systems. 209
5.4.16         The Reserved Variable ?self.. 209
5.4.17         Virtual Environment 210
5.5      Design of Conceptual Processes. 211
5.5.1       TalaMind Control Flow.. 211
5.5.2       Design of Executable Concepts. 214
5.5.3       Pattern-Matching. 217
5.5.4       Tala Constructions. 218
5.5.5       Tala Processing of Goals. 222
5.6      Design of User Interface. 223
5.6.1       Design of the TalaMind Applet 223
5.6.2       FlatEnglish Display. 227
5.7      Summary. 228
6.        Demonstration. 230
6.1      Overview.. 230
6.2      Demonstration Content 231
6.2.1       The Discovery of Bread Story Simulation. 231
6.2.2       The Farmer’s Dilemma Story Simulation. 234
6.3      Illustration of Higher-Level Mentalities. 236
6.3.1       Natural Language Understanding. 236
6.3.2       Multi-Level Reasoning. 237
6.3.2.1     Deduction. 237
6.3.2.2     Induction. 237
6.3.2.3     Abduction, Analogy, Causality, Purpose. 238
6.3.2.4     Meta-Reasoning. 239
6.3.3       Self-Development and Higher-Level Learning. 240
6.3.3.1     Analogy, Causality, and Purpose in Learning. 240
6.3.3.2     Learning by Reflection and Self-Programming. 240
6.3.3.3     Learning by Invention of Languages. 241
6.3.4       Curiosity. 241
6.3.5       Imagination. 241
6.3.5.1     Imagination via Conceptual Blends. 242
6.3.5.2     Imagination via Nested Conceptual Simulation. 244
6.3.6       Consciousness. 246
6.4      Summary. 247
7.        Evaluation. 248
7.1      Criteria for Evaluating Plausibility. 248
7.2      Theoretical Issues and Objections. 248
7.3      Affirmative Theoretical Arguments. 249
7.4      Design and Demonstration. 250
7.5   Novelty in Relation to Previous Research   251
Figures. viii
§ Notation and Overview of Changes. ix
Synopsis. xi
Preface. xiii
1.        Introduction. 1
1.1      Can Machines Have Human-Level Intelligence?. 1
1.2      Thesis Approach. 4
1.3      Terminology: Tala and TalaMind. 7
1.4      TalaMind Hypotheses. 7
1.4.1       Intelligence Kernel Hypothesis. 8
1.4.2       Natural Language Mentalese Hypothesis. 9
1.4.3       Multiple Levels of Mentality Hypothesis. 11
1.4.4       Relation to the Physical Symbol System Hypothesis. 11
1.5      TalaMind System Architecture. 12
1.6      Arguments and Evidence: Strategy and Criteria for Success. 16
1.7      Overview of Chapters. 18
2.        Subject Review: Human-Level AI and Natural Language. 19
2.1      Human-Level Artificial Intelligence. 19
2.1.1       How to Define and Recognize Human-Level AI 19
2.1.2       Unexplained Features of Human-Level Intelligence. 22
2.1.2.1     Generality. 22
2.1.2.2     Creativity and Originality. 23
2.1.2.3     Natural Language Understanding. 24
2.1.2.4     Effectiveness, Robustness, Efficiency. 24
2.1.2.5     Self-Development and Higher-Level Learning. 24
2.1.2.6     Metacognition and Multi-Level Reasoning. 25
2.1.2.7     Imagination. 26
2.1.2.8     Consciousness. 26
2.1.2.9     Sociality, Emotions, Values. 27
2.1.2.10     Visualization, Spatial-Temporal Reasoning. 28
2.1.2.11     Curiosity, Self-Programming, Theory of Mind. 28
2.1.2.12     Other Unexplained Features. 29
2.2      Natural Language. 29
2.2.1       Does Thought Require Language?. 29
2.2.2       What Does Meaning Mean?. 33
2.2.3       Does Human-Level AI Require Embodiment?. 37
2.2.4       Natural Language, Metacognition, Inner Speech. 39
2.3      Relation of Thesis Approach to Previous Research. 41
2.3.1       Formal, Logical Approaches. 41
2.3.2       Cognitive Approaches and Cognitive Linguistics. 42
2.3.3       Approaches to Human-Level Artificial Intelligence. 45
2.3.3.1     Sloman. 45
2.3.3.2     Minsky. 45
2.3.3.3     McCarthy. 47
2.3.3.4     Reverse-Engineering the Brain. 48
2.3.3.5     Cognitive Architectures and AGI 49
2.3.3.6     Newell and Simon’s Cognitive Research. 50
2.3.3.6.1     Unified Theories of Cognition. 50
2.3.3.6.2     The ‘Knowledge Level’ and ‘Intelligence Level’ 51
2.3.3.7     Other Influences for Thesis Approach. 53
2.3.4       Approaches to Artificial Consciousness. 53
2.3.5       Approaches to Reflection and Self-Programming. 55
2.3.6       Johnson-Laird’s Mental Models. 59
2.3.7       Research on Natural Logic. 61
2.3.7.1     Natural Logic According to Lakoff 61
2.3.7.2     Monotonicity-Based Natural Logic. 62
2.4      Summary. 64
3.        Analysis of Thesis Approach to Human-Level AI 65
3.1      Overview.. 65
3.2      Theoretical Requirements for TalaMind Architecture. 66
3.2.1       Conceptual Language. 66
3.2.2       Conceptual Framework. 70
3.2.3       Conceptual Processes. 72
3.3      Representing Meaning with Natural Language Syntax. 73
3.4      Representing English Syntax in Tala. 76
3.4.1       Non-Prescriptive, Open, Flexible. 76
3.4.2       Semantic and Ontological Neutrality and Generality. 77
3.5      Choices and Methods for Representing English Syntax. 77
3.5.1       Theoretical Approach to Represent English Syntax. 77
3.5.2       Representing Syntactic Structure of NL Sentences. 78
3.6      Semantic Representation and Processing. 80
3.6.1       Lexemes, Senses, Referents, and Variables. 80
3.6.2       Multiple Representations for the Same Concept 83
3.6.3       Representing Interpretations. 84
3.6.3.1     Underspecification. 85
3.6.3.2     Syntactic Elimination of Interpretations. 85
3.6.3.3     Generic and Non-Generic Interpretations. 87
3.6.3.4     Specific and Non-Specific Interpretations. 87
3.6.3.5     Individual and Collective Interpretations. 87
3.6.3.6     Count and Mass Interpretations. 87
3.6.3.7     Quantificational Interpretations. 88
3.6.3.8     De Dicto and De Re Interpretations. 90
3.6.3.9     Interpretations of Compound Noun Structures. 92
3.6.3.10     Interpretations of Metaphors. 93
3.6.3.11     Interpretations of Metonyms. 93
3.6.3.12     Interpretations of Anaphora. 93
3.6.3.13     Interpretation of Idioms. 94
3.6.4       Semantic Disambiguation. 95
3.6.5       Representing Implications. 95
3.6.6       Semantic Inference. 96
3.6.6.1     Representation of Truth. 97
3.6.6.2     Negation and Contradictions. 97
3.6.6.3     Inference with Commonsense. 101
3.6.6.4     Paraphrase and Inference. 102
3.6.6.5     Inference for Metaphors and Metonyms. 102
3.6.7       Representation of Contexts. 103
3.6.7.1     Dimensions of Context 103
3.6.7.2     Perceived Reality. 107
3.6.7.3     Event Memory. 107
3.6.7.4     Encyclopedic and Commonsense Knowledge. 107
3.6.7.5     Interactive Contexts and Mutual Knowledge. 109
3.6.7.6     Hypothetical Contexts. 112
3.6.7.7     Semantic Domains. 113
3.6.7.8     Mental Spaces. 114
3.6.7.9     Conceptual Blends. 118
3.6.7.10     Theory Contexts. 121
3.6.7.11     Problem Contexts. 123
3.6.7.12     Composite Contexts. 124
3.6.7.13     Society of Mind Thought Context 124
3.6.7.14     Meta-Contexts. 124
3.6.8       Primitive Words and Variables in Tala. 125
3.7      Higher-Level Mentalities. 128
3.7.1       Multi-Level Reasoning. 129
3.7.1.1     Deduction. 129
3.7.1.2     Induction. 129
3.7.1.3     Abduction. 129
3.7.1.4     Analogical Reasoning. 130
3.7.1.5     Causal and Purposive Reasoning. 130
3.7.1.6     Meta-Reasoning. 131
3.7.2       Self-Development and Higher-Level Learning. 131
3.7.2.1     Learning by Multi-Level Reasoning. 131
3.7.2.2     Learning by Reflection and Self-Programming. 132
3.7.2.3     Learning by Invention of Languages. 134
3.7.3       Curiosity. 135
3.7.4       Imagination. 138
3.7.5       Sociality, Emotions, Values. 139
3.7.6       Consciousness. 139
3.8      Summary. 141
4.        Theoretical Issues and Objections. 142
4.1      Issues and Objections re the Possibility of Human-Level AI 142
4.1.1       Dreyfus Issues. 142
4.1.2       Penrose Objections. 144
4.1.2.1     General Claims re Intelligence. 144
4.1.2.2     Claims re Human Logical Insight 146
4.1.2.3     Gödelian Arguments. 147
4.1.2.4     Continuous Computation. 154
4.1.2.5     Hypothesis re Orchestrated Objective Reduction. 154
4.2      Issues and Objections for Thesis Approach. 156
4.2.1       Theoretical Objections to a Language of Thought 156
4.2.2       Objections to Representing Semantics via NL Syntax. 156
4.2.2.1     The Circularity Objection. 156
4.2.2.2     Objection Syntax Is Insufficient for Semantics. 156
4.2.2.3     Ambiguity Objections to Natural Language. 157
4.2.2.4     Objection Thought Is Perceptual, Not Linguistic. 158
4.2.3       Weizenbaum’s Eliza Program.. 159
4.2.4       Searle’s Chinese Room Argument 161
4.2.5       McCarthy’s Objections to Natural Language Mentalese. 164
4.2.6       Minsky’s Issues for Representation and Learning. 167
4.2.7       Chalmers’ Hard Problem of Consciousness. 169
4.2.8       Smith’s Issues for Representation and Reflection. 172
4.3      Summary. 177
5.        Design of a Demonstration System... 178
5.1      Overview.. 178
5.2      Nature of the Demonstration System.. 179
5.3      Design of Conceptual Language. 180
5.3.1       Tala Syntax Notation. 181
5.3.2       Nouns. 182
5.3.3       Verbs. 185
5.3.4       Prepositions. 188
5.3.5       Pronouns. 190
5.3.6       Determiners. 190
5.3.7       Adjectives. 192
5.3.8       Adverbs. 193
5.3.9       Conjunctions. 193
5.3.9.1     Coordinating Conjunctions. 194
5.3.9.2     Subordinating / Structured Conjunctions. 196
5.3.9.3     Correlative Conjunctions. 198
5.3.10         Interjections. 199
5.3.11         Tala Variables and Pointers. 199
5.3.12         Inflections. 199
5.3.12.1     Determiner-Complement Agreement 200
5.3.12.2     Subject-Verb Agreement 200
5.4      Design of Conceptual Framework. 201
5.4.1       Requirements for a Conceptual Framework. 201
5.4.2       Structure of the Conceptual Framework. 202
5.4.3       Perceived Reality – Percepts and Effepts. 204
5.4.4       Subagents, Mpercepts, and Meffepts. 205
5.4.5       Tala Lexicon. 205
5.4.6       Encyclopedic Knowledge and Semantic Domains. 206
5.4.7       Current Domains. 207
5.4.8       Mental Spaces and Conceptual Blends. 207
5.4.9       Scenarios. 207
5.4.10         Thoughts. 208
5.4.11         Goals. 208
5.4.12         Executable Concepts. 208
5.4.13         Tala Constructions and Metaphors. 209
5.4.14         Event-Memory. 209
5.4.15         Systems. 209
5.4.16         The Reserved Variable ?self.. 209
5.4.17         Virtual Environment 210
5.5      Design of Conceptual Processes. 211
5.5.1       TalaMind Control Flow.. 211
5.5.2       Design of Executable Concepts. 214
5.5.3       Pattern-Matching. 217
5.5.4       Tala Constructions. 218
5.5.5       Tala Processing of Goals. 222
5.6      Design of User Interface. 223
5.6.1       Design of the TalaMind Applet 223
5.6.2       FlatEnglish Display. 227
5.7      Summary. 228
6.        Demonstration. 230
6.1      Overview.. 230
6.2      Demonstration Content 231
6.2.1       The Discovery of Bread Story Simulation. 231
6.2.2       The Farmer’s Dilemma Story Simulation. 234
6.3      Illustration of Higher-Level Mentalities. 236
6.3.1       Natural Language Understanding. 236
6.3.2       Multi-Level Reasoning. 237
6.3.2.1     Deduction. 237
6.3.2.2     Induction. 237
6.3.2.3     Abduction, Analogy, Causality, Purpose. 238
6.3.2.4     Meta-Reasoning. 239
6.3.3       Self-Development and Higher-Level Learning. 240
6.3.3.1     Analogy, Causality, and Purpose in Learning. 240
6.3.3.2     Learning by Reflection and Self-Programming. 240
6.3.3.3     Learning by Invention of Languages. 241
6.3.4       Curiosity. 241
6.3.5       Imagination. 241
6.3.5.1     Imagination via Conceptual Blends. 242
6.3.5.2     Imagination via Nested Conceptual Simulation. 244
6.3.6       Consciousness. 246
6.4      Summary. 247
7.        Evaluation. 248
7.1      Criteria for Evaluating Plausibility. 248
7.2      Theoretical Issues and Objections. 248
7.3      Affirmative Theoretical Arguments. 249
7.4      Design and Demonstration. 250
7.5      Novelty in Relation to Previous Research. 251
7.6      Areas for Future AI Research. 252
7.7      Plausibility of Thesis Approach. 254
8.        Future Potentials. 256
8.1      Potential Economic Consequences. 257
8.2      Toward Beneficial Human-Level AI and Superintelligence. 260
8.2.1       Importance of TalaMind for Beneficial AI 261
8.2.2       AI’s Different Concept of Self-Preservation. 261
8.2.3       Symbolic Consciousness ≠ Human Consciousness. 262
8.2.4       A Counter-Argument Invoking PSSH.. 262
8.2.5       Acting As If Robots Are Fully Conscious. 263
8.2.6       Avoiding Artificial Slavery. 263
8.2.7       Theory of Mind and Simulations of Minds. 264
8.2.8       A Mind Is a Universe Unto Itself 264
8.2.9       Uploading Human Consciousness. 265
8.2.10         The Possibility of Superintelligence. 266
8.2.11         Completeness of Human Intelligence. 268
8.2.12         Nature of Thought and Conceptual Gaps. 269
8.2.13         Is ‘Strong' Superintelligence Possible?. 269
8.2.14         Two Paths to Superintelligence. 270
8.2.15         Human-Level AI and Goals. 271
8.2.16         TalaMind’s Role in Beneficial Superintelligence. 272
8.2.17         Future Challenges for Human-Level AI+ via TalaMind. 272
8.2.18         When Will Human-Level AI Be Achieved?. 273
8.3      Humanity’s Long-Term Prosperity and Survival 273
9.        Summation. 275
Glossary. 278
Appendix A. Theoretical Questions for Analysis of Approach. 283
Appendix B. Processing in Discovery of Bread Simulation. 286
Bibliography. 310