Computers

Preface

The book’s Preface places the current AI explosion in the context of other technological cataclysms and recounts the author’s personal (and not always deadly serious) AI journey.

Chapter One: “Extracting the Essence: Toward Machine Translation of Literature” 

This 2019 essay rashly inquired whether artificial intelligence (AI) and machine translation (MT) might eventually be applied to literary translation. Such translation strives to somehow preserve the essence of a work while carrying it over to a different language and culture and giving it rebirth there. To recognize that essence, the translator must accurately capture the meaning of the original; appreciate its metaphors, connotations, register, references, and other abstract or associative factors; and choose among available target language expressions by exercising esthetic judgments. Computers, however, have until recently remained incapable of such accuracy, abstraction, and judgment. We revisited these shortfalls in light of developments in MT and AI. We teased apart several separable aspects of literary translation – literal meaning, meter, rhyme, and the abovementioned associative elements – with reference to arguments about Vladimir Nabokov’s hyper-literal translation of Pushkin’s poem Eugene Onegin. Prompted by this debate, we came to analyze translation as an optimization problem: because it will often prove impossible to perfectly convey all aspects or essences of a text in a single translation, the translator must search for some optimal compromise. Then we discussed several avenues for improvement in MT which could help to extract these aspects of a text’s essence – first, those which might enhance textually grounded MT (i.e., MT trained on text only), leading to delivery of high-quality literal translations; and second, those related to perceptually grounded MT (i.e., MT trained on simulated perception, e.g. of audiovisual input, as well as text), which might extract more abstract or associative elements of a text. We suggested that recognition of categories would prove central to the essence extraction sought by translators. As this categorization improves, MT should increasingly support literary, and thus cultural, preservation. However, the deepest artificial esthetic judgments will await artificial emotion. Chapter One concluded with two appendices, the first sampling numerous competing translations of Pushkin verses and the second displaying widely varying translations of a short French poem.

Chapter Two: “Toward an Artificial Nabokov”

Chapter One inquired whether artificial translation of literature might be at all possible. The cautious conclusion was that while high-quality literal translation might be achievable through foreseeable development of current techniques, artistic translation would await artificial emotion, a more distant prospect. This 2021 sequel went on to ask whether an artificial intelligence might eventually gain the ability to actually create works of literary art. To throw literary consciousness into the sharpest relief, we took as exemplar an author known for a kind of hyper-consciousness: Vladimir Nabokov. To be sure, the suggestion that artworks combining Nabokov’s superhuman intricacy and wholly human depth could be authored by a collection of switches would horrify this transcendent author, and does seem to fly in the face of everything that is most human. But while we are concerned with what machines might do, our more fundamental concern is to understand the human thoughts and feelings to which machines might aspire; and this understanding, promising to bridge the gap between C.P. Snow’s two cultures, is finally coming within reach. In our literary context, Nabokov scholarship provides many specific examples – in Ada: or Ardor, Pale Fire, and other works – of the author’s hyper-conscious artistic techniques: glorying in memory; repetition to establish themes and motifs; allusion to wide-ranging works and facts; intricate puzzle posing; and relentlessly careful structuring at multiple levels of the text. Here we considered several such techniques, speculating about the extent to which current or coming AI capabilities could approach them. In Chapter Two, Section 2, to clarify assumptions, I set forth my own current conceptions of computation, consciousness, feeling, language, and thinking, providing in the process a somewhat prejudiced AI primer for the computer-shy humanist. In Chapter Two, Section 3, I applied to Nabokov’s prodigious work my understanding of these aspects of mind. Subsections focused on self-awareness, perception, memory, and puzzles.

Chapter Three: “Large Literary Models? Intelligence and Language in the LLM Era”

Spoiler: Chapter Three, Section 6, Experiments will present several striking demonstrations of the current (2025) state of the artificial literary art. This chapter’s initial sections aim to explain the breakthroughs that triggered the abrupt phase change from wannabe to indisputable intelligence and linguistic ability. We first give an account of intelligence, sufficiently general to apply to both biological and artificial entities, defining it as the ability to select actions or outcomes effectively according to the conditions and goals encountered – in computers, as conditional (if/then) expressions. We'll scan various ways of packaging conditionals in computer programs, culminating in deep neural network technology, in which each network node among billions can be seen as an if/then expression. Thus, conditionality realized through networks is seen as the common underpinning of artificial and biological intelligence. Next, to explain the breakthrough success of Large Language Models, we undertake an accompanying account of language, viewed as combining two separable capabilities: (1) to communicate using symbols, minimally one at a time and (2) to communicate with a sequence of symbols – that is, exploiting grammar. We explain the technological breakthrough enabling artificial symbol use as development of vector-based semantic techniques and we explain in depth our understanding of symbolic communication. In the grammatical area, the linguistic breakthrough has been enablement of improved predictors of sequences through consideration of much larger contexts. To manage the accompanying threat of computational overload, it’s necessary to focus on the most predictive contextual elements among thousands. These are taken to be those closest semantically to a given element in question – and the vector-based semantic approach proves to be perfect for identifying them via the attention mechanism and the sequence-prediction technology built upon it, the transformer architecture. However, despite spectacular progress in computational intelligence and language, some aspects undoubtedly remain lacking. We emphasize that the structure of neurally learned knowledge remains unclear, while speculating that class hierarchies and schemas play important parts in LLMs and describing new tools for analyzing the networks’ hidden patterns. We go on to consider issues of (1) search and revision (linguistic and otherwise); (2) experience grounded in the world beyond text, with explanation of Imitation Learning and Reinforcement Learning as ways of predicting what to do next; (3) memory and identity issues; and (4) emotions, stressing that current systems’ lack of built-in drives handicaps artificial artistry – for better or worse – and cautioning against mistaking faked feelings for felt ones. After presenting the abovementioned experiments, we'll conclude with an evaluation, attempting a working definition of “understanding” to support the contention that current LLMs do in fact evince artificial comprehension, while also noting original touches and linguistic creativity.

Sendoff

The volume’s Sendoff considers whether machines, while now arguably intelligent, can ever also gain sufficient sensation and emotion to create language art with other than borrowed depth – and, while viewing this development as likely all too soon, declines to despair on that account.

Preface

The book’s Preface places the current AI explosion in the context of other technological cataclysms and recounts the author’s personal (and not always deadly serious) AI journey.

Chapter One: “Extracting the Essence: Toward Machine Translation of Literature” 

This 2019 essay rashly inquired whether artificial intelligence (AI) and machine translation (MT) might eventually be applied to literary translation. Such translation strives to somehow preserve the essence of a work while carrying it over to a different language and culture and giving it rebirth there. To recognize that essence, the translator must accurately capture the meaning of the original; appreciate its metaphors, connotations, register, references, and other abstract or associative factors; and choose among available target language expressions by exercising esthetic judgments. Computers, however, have until recently remained incapable of such accuracy, abstraction, and judgment. We revisited these shortfalls in light of developments in MT and AI. We teased apart several separable aspects of literary translation – literal meaning, meter, rhyme, and the abovementioned associative elements – with reference to arguments about Vladimir Nabokov’s hyper-literal translation of Pushkin’s poem Eugene Onegin. Prompted by this debate, we came to analyze translation as an optimization problem: because it will often prove impossible to perfectly convey all aspects or essences of a text in a single translation, the translator must search for some optimal compromise. Then we discussed several avenues for improvement in MT which could help to extract these aspects of a text’s essence – first, those which might enhance textually grounded MT (i.e., MT trained on text only), leading to delivery of high-quality literal translations; and second, those related to perceptually grounded MT (i.e., MT trained on simulated perception, e.g. of audiovisual input, as well as text), which might extract more abstract or associative elements of a text. We suggested that recognition of categories would prove central to the essence extraction sought by translators. As this categorization improves, MT should increasingly support literary, and thus cultural, preservation. However, the deepest artificial esthetic judgments will await artificial emotion. Chapter One concluded with two appendices, the first sampling numerous competing translations of Pushkin verses and the second displaying widely varying translations of a short French poem.

Chapter Two: “Toward an Artificial Nabokov”

Chapter One inquired whether artificial translation of literature might be at all possible. The cautious conclusion was that while high-quality literal translation might be achievable through foreseeable development of current techniques, artistic translation would await artificial emotion, a more distant prospect. This 2021 sequel went on to ask whether an artificial intelligence might eventually gain the ability to actually create works of literary art. To throw literary consciousness into the sharpest relief, we took as exemplar an author known for a kind of hyper-consciousness: Vladimir Nabokov. To be sure, the suggestion that artworks combining Nabokov’s superhuman intricacy and wholly human depth could be authored by a collection of switches would horrify this transcendent author, and does seem to fly in the face of everything that is most human. But while we are concerned with what machines might do, our more fundamental concern is to understand the human thoughts and feelings to which machines might aspire; and this understanding, promising to bridge the gap between C.P. Snow’s two cultures, is finally coming within reach. In our literary context, Nabokov scholarship provides many specific examples – in Ada: or Ardor, Pale Fire, and other works – of the author’s hyper-conscious artistic techniques: glorying in memory; repetition to establish themes and motifs; allusion to wide-ranging works and facts; intricate puzzle posing; and relentlessly careful structuring at multiple levels of the text. Here we considered several such techniques, speculating about the extent to which current or coming AI capabilities could approach them. In Chapter Two, Section 2, to clarify assumptions, I set forth my own current conceptions of computation, consciousness, feeling, language, and thinking, providing in the process a somewhat prejudiced AI primer for the computer-shy humanist. In Chapter Two, Section 3, I applied to Nabokov’s prodigious work my understanding of these aspects of mind. Subsections focused on self-awareness, perception, memory, and puzzles.

Chapter Three: “Large Literary Models? Intelligence and Language in the LLM Era”

Spoiler: Chapter Three, Section 6, Experiments will present several striking demonstrations of the current (2025) state of the artificial literary art. This chapter’s initial sections aim to explain the breakthroughs that triggered the abrupt phase change from wannabe to indisputable intelligence and linguistic ability. We first give an account of intelligence, sufficiently general to apply to both biological and artificial entities, defining it as the ability to select actions or outcomes effectively according to the conditions and goals encountered – in computers, as conditional (if/then) expressions. We'll scan various ways of packaging conditionals in computer programs, culminating in deep neural network technology, in which each network node among billions can be seen as an if/then expression. Thus, conditionality realized through networks is seen as the common underpinning of artificial and biological intelligence. Next, to explain the breakthrough success of Large Language Models, we undertake an accompanying account of language, viewed as combining two separable capabilities: (1) to communicate using symbols, minimally one at a time and (2) to communicate with a sequence of symbols – that is, exploiting grammar. We explain the technological breakthrough enabling artificial symbol use as development of vector-based semantic techniques and we explain in depth our understanding of symbolic communication. In the grammatical area, the linguistic breakthrough has been enablement of improved predictors of sequences through consideration of much larger contexts. To manage the accompanying threat of computational overload, it’s necessary to focus on the most predictive contextual elements among thousands. These are taken to be those closest semantically to a given element in question – and the vector-based semantic approach proves to be perfect for identifying them via the attention mechanism and the sequence-prediction technology built upon it, the transformer architecture. However, despite spectacular progress in computational intelligence and language, some aspects undoubtedly remain lacking. We emphasize that the structure of neurally learned knowledge remains unclear, while speculating that class hierarchies and schemas play important parts in LLMs and describing new tools for analyzing the networks’ hidden patterns. We go on to consider issues of (1) search and revision (linguistic and otherwise); (2) experience grounded in the world beyond text, with explanation of Imitation Learning and Reinforcement Learning as ways of predicting what to do next; (3) memory and identity issues; and (4) emotions, stressing that current systems’ lack of built-in drives handicaps artificial artistry – for better or worse – and cautioning against mistaking faked feelings for felt ones. After presenting the abovementioned experiments, we'll conclude with an evaluation, attempting a working definition of “understanding” to support the contention that current LLMs do in fact evince artificial comprehension, while also noting original touches and linguistic creativity.

Sendoff

The volume’s Sendoff considers whether machines, while now arguably intelligent, can ever also gain sufficient sensation and emotion to create language art with other than borrowed depth – and, while viewing this development as likely all too soon, declines to despair on that account.

On the commercial side, artificial intelligence applications are powering many sectors. Globally, governments are exploring how to comprehend, incorporate, apply, and use artificial intelligence technologies. The scope of government use of artificial intelligence technology goes beyond that of commercial organizations and is far more complex. In government, the challenges will be as follows: (1) How can governments use artificial intelligence technology to improve their efficiencies? (2) How can governments become more citizen-centric, service based, accessible, and responsive? (3) How can governments protect their citizens from the misuse of artificial intelligence  (e.g., alleged Russian bots’ interference in U.S. elections)? (4)  How can governments use artificial intelligence technology to make better policy decisions and avoid wrong decisions  (economic, social etc.)? (5) How can governments develop new standards to govern and manage the deployment of artificial intelligence technologies (e.g., autonomous cars, financial markets and trading, healthcare bots)? (6) How will the legislative bodies respond to the rise of intelligent machines? (7) How will the use of artificial intelligence in the military change the arms race? (8)  What roles governments will need to play in developing global standards related to artificial intelligence (United Nations)? (9) How can governments improve their countries’ productivity with artificial intelligence? (10) How can governments handle the upcoming unemployment that would result from AI automation? All the above questions are at an early stage of exploration and many have not been addressed comprehensively. This book deals with all the above issues and provides the first guide to governments and policy makers of the world on artificial intelligence.

Digital technologies are an indispensable facet of every aspect of our society. Even in the developing world, mobile phones have transformed the lives and livelihoods of average citizens. Yet, two decades ago, when there were more phone lines in Manhattan than in most of Sub-Saharan Africa, only a few visionary institutions could have imagined that computers, the Internet and mobile phones would be so prominent in poverty-stricken environments. One of these visionary institutions was the International Development Research Centre (IDRC), which recognized the important but complex role that information and communication technologies (ICTs) would have in fostering human development and reducing poverty. IDRC-supported projects critically examined the ways in which ICTs could be used to improve learning, empower the disenfranchised, generate income opportunities for the poor, and facilitate access to healthcare in Africa, Asia, Latin America and the Caribbean. Their research focused on development priorities that were defined in collaboration with researchers from the Global South, civil society organizations, government officials and policymakers. By supporting research in this field since 1996, IDRC has become one of the leading institutions and key contributors to the growth of the “ICTs for development” (ICT4D) field, specifically because of its strategic decision to focus on building the capacity of Southern researchers and policymakers to explore how ICTs can continue to change people’s lives in the developing world.

Considering that most development institutions and governments are currently attempting to integrate ICTs into their practices, this is an opportune time to reflect on the research findings that have emerged from working alongside researchers in this area. In particular, this book examines how research has helped IDRC contribute to building the ICT4D field based on a nuanced understanding of the relationship between ICTs and development goals. It also discusses programmatic investments made by IDRC since the late 1990s in a wide variety of areas related to ICTs, including infrastructure, access, regulations, health, governance, education, livelihoods, social inclusion, technical innovation, intellectual property rights and evaluation.

Each chapter in this book analyzes how the research findings from IDRC-supported projects have contributed to an evolution of thinking, and the successes and challenges in using ICTs as a tool to address development issues. Each chapter also presents key lessons learned from ICT4D programming and makes recommendations for future work. The book illustrates how IDRC’s focus shifted over time from looking specifically at issues of access to understanding the implications of ICTs in the lives of citizens in the developing world. 

Digital technologies are an indispensable facet of every aspect of our society. Even in the developing world, mobile phones have transformed the lives and livelihoods of average citizens. Yet, two decades ago, when there were more phone lines in Manhattan than in most of Sub-Saharan Africa, only a few visionary institutions could have imagined that computers, the Internet and mobile phones would be so prominent in poverty-stricken environments. One of these visionary institutions was the International Development Research Centre (IDRC), which recognized the important but complex role that information and communication technologies (ICTs) would have in fostering human development and reducing poverty. IDRC-supported projects critically examined the ways in which ICTs could be used to improve learning, empower the disenfranchised, generate income opportunities for the poor, and facilitate access to healthcare in Africa, Asia, Latin America and the Caribbean. Their research focused on development priorities that were defined in collaboration with researchers from the Global South, civil society organizations, government officials and policymakers. By supporting research in this field since 1996, IDRC has become one of the leading institutions and key contributors to the growth of the “ICTs for development” (ICT4D) field, specifically because of its strategic decision to focus on building the capacity of Southern researchers and policymakers to explore how ICTs can continue to change people’s lives in the developing world.

Considering that most development institutions and governments are currently attempting to integrate ICTs into their practices, this is an opportune time to reflect on the research findings that have emerged from working alongside researchers in this area. In particular, this book examines how research has helped IDRC contribute to building the ICT4D field based on a nuanced understanding of the relationship between ICTs and development goals. It also discusses programmatic investments made by IDRC since the late 1990s in a wide variety of areas related to ICTs, including infrastructure, access, regulations, health, governance, education, livelihoods, social inclusion, technical innovation, intellectual property rights and evaluation.

Each chapter in this book analyzes how the research findings from IDRC-supported projects have contributed to an evolution of thinking, and the successes and challenges in using ICTs as a tool to address development issues. Each chapter also presents key lessons learned from ICT4D programming and makes recommendations for future work. The book illustrates how IDRC’s focus shifted over time from looking specifically at issues of access to understanding the implications of ICTs in the lives of citizens in the developing world. 

This book explores the history of hypertext, an influential concept that forms the underlying structure of the World Wide Web and innumerable software applications. Barnet combines an analysis of contemporary literature with her exclusive interviews with those at the forefront of the hypertext innovation. She tells both the human and the technological story, tracing its path back to an analogue device imagined by Vannevar Bush in 1945, before modern computing had happened.

‘Memory Machines’ offers an expansive record of hypertext over the last 60 years, pinpointing the major breakthroughs and fundamental flaws in its evolution. Barnet argues that some of the earliest hypertext systems were more richly connected and in some respects more flexible than the Web; this is also a fascinating account of the paths not taken.

Barnet ends the journey through computing history at the birth of mass domesticated hypertext, at the point that it grew out of the university labs and into the Web. And yet she suggests that hypertext may not have completed its evolutionary story, and may still have the capacity to become something different, something much better than it is today.

This book explores the history of hypertext, an influential concept that forms the underlying structure of the World Wide Web and innumerable software applications. Barnet combines an analysis of contemporary literature with her exclusive interviews with those at the forefront of the hypertext innovation. She tells both the human and the technological story, tracing its path back to an analogue device imagined by Vannevar Bush in 1945, before modern computing had happened.

‘Memory Machines’ offers an expansive record of hypertext over the last 60 years, pinpointing the major breakthroughs and fundamental flaws in its evolution. Barnet argues that some of the earliest hypertext systems were more richly connected and in some respects more flexible than the Web; this is also a fascinating account of the paths not taken.

Barnet ends the journey through computing history at the birth of mass domesticated hypertext, at the point that it grew out of the university labs and into the Web. And yet she suggests that hypertext may not have completed its evolutionary story, and may still have the capacity to become something different, something much better than it is today.

Unfortunately, developing countries and less developed countries in general have not yet entered the digital era. Most of them have not yet developed the back-office components that are fundamental prerequisites for conducting e-applications. In many situations, e-government systems have been adopted solely as window dressing, as it is considered improper for governmental agencies not to have a web portal, email address and/or a Facebook or Twitter account. But these government web portals are of no real use to the citizens. This volume seeks to help rectify this issue.

Drawing lessons from the eFez Project in Morocco, “E-Government for Good Governance in Developing Countries” offers practical supporting material to decision makers in developing countries on information and communication technologies for development (ICT4D), specifically e-government implementation. It documents the eFez Project experience in all of its aspects, presenting the project’s findings and the practical methods developed by the authors (a roadmap, impact assessment framework, design issues, lessons learned and best practices) in their systematic quest to turn eFez’s indigenous experimentations and findings into a formal framework for academics, practitioners and decision makers. The volume also reviews, analyzes and synthesizes the findings of other projects to offer a comparative study of the eFez framework and a number of other e-government frameworks from the growing literature.

Given the lack of practical books that target decision makers guiding the design and implementation of e-government for good governance and any other sector-specific ICT4D, the authors hope that the eFez Project’s great success in Morocco, and the outcomes and methods described in this volume, will prove a useful model for practitioners and decision makers in other developing countries around the world. 

Unfortunately, developing countries and less developed countries in general have not yet entered the digital era. Most of them have not yet developed the back-office components that are fundamental prerequisites for conducting e-applications. In many situations, e-government systems have been adopted solely as window dressing, as it is considered improper for governmental agencies not to have a web portal, email address and/or a Facebook or Twitter account. But these government web portals are of no real use to the citizens. This volume seeks to help rectify this issue.

Drawing lessons from the eFez Project in Morocco, “E-Government for Good Governance in Developing Countries” offers practical supporting material to decision makers in developing countries on information and communication technologies for development (ICT4D), specifically e-government implementation. It documents the eFez Project experience in all of its aspects, presenting the project’s findings and the practical methods developed by the authors (a roadmap, impact assessment framework, design issues, lessons learned and best practices) in their systematic quest to turn eFez’s indigenous experimentations and findings into a formal framework for academics, practitioners and decision makers. The volume also reviews, analyzes and synthesizes the findings of other projects to offer a comparative study of the eFez framework and a number of other e-government frameworks from the growing literature.

Given the lack of practical books that target decision makers guiding the design and implementation of e-government for good governance and any other sector-specific ICT4D, the authors hope that the eFez Project’s great success in Morocco, and the outcomes and methods described in this volume, will prove a useful model for practitioners and decision makers in other developing countries around the world.