Scholarly research and writing requires establishing context and importance for readers at the very beginning. A well-crafted introduction aims to introduce the topic, provide background information to orient readers, convey the significance of the subject matter, raise key questions or issues to be examined, and provide an overview or preview of what is to come in the paper. This introduction strives to accomplish all of these goals for a hypothetical research paper analyzing the linguistic and societal impacts of code-switching among bilingual communities in the United States.
Code-switching, which refers to alternating between two or more languages or language varieties in conversation, is a common phenomenon among multilingual populations worldwide but one that remains inadequately documented and understood, especially in the American context. As globalization increases interaction and immigration draws more language minorities to the U.S., the prevalence of bilingualism and code-switching is rising dramatically. The latest census data indicates that over 60 million Americans speak a language other than English at home, with Spanish, Chinese, Tagalog, Vietnamese, Arabic and several other languages regularly code-switched in public and private communications (US Census Bureau, 2015). Relatively few empirical studies have systematically analyzed examples of code-switching in American bilingual communities to discern linguistic patterns and meanings, or contextualized such linguistic behaviors within their broader social functions and implications.
Scholars across diverse fields have begun calling more attention to code-switching as an important area meriting increased focus and interdisciplinary research. Linguists seek to better comprehend the structural mechanisms, rule systems and pragmatic cues that govern code-switching, while sociologists, anthropologists and communication experts investigate the cultural motivations and social impacts. As Myers-Scotton (1993) aptly noted, “Codeswitching provides a window on the relationship between language and social structure” (p. 1). Examining naturally occurring code-switches can reveal much about a bilingual group’s values, power dynamics, and negotiation of ethnic and national identities. At the same time, psychologists have an interest in how code-switching relates to concepts of self, cognitive processing, and language acquisition/attrition.
Recent studies suggest code-switching fills important interactional roles even beyond matters of personal identity. Findings imply it serves communicative functions like emphasis, humor, swearing, conveying emotion and managing social relationships (Bullock & Toribio, 2009). For these reasons, educational researchers are exploring code-switching’s implications for bilingual instruction and academic achievement. Politicians and policymakers debate whether permissive or restrictive stances on code-switching better promote social cohesion or cultural maintenance among immigrant groups. The news media and general public also exhibit varied perceptions of code-switching as a sign of linguistic deficiency, cultural pride or hybrid identity formation.
As the preceding overview illuminates, code-switching constitutes a rich site of interdisciplinary inquiry with significant theoretical, practical and societal consequences remaining to be uncovered. Much existing research focuses on code-switching in European, Asian or African linguistic contexts rather than the American experience. Therefore, this paper aims to address key gaps in the literature and furtherknowledge on this topic through an in-depth examination of code-switching patterns, functions and implications among selected bilingual groups residing in the United States. The following section outlines this research’s specific purpose and guiding questions. Through analysis of quantitative survey data and qualitative interviews, it seeks to deliver valuable new insights into the linguistic structures and sociocultural dimensions of code-switching as exercised by Spanish-English and Chinese-English bilinguals in America today. Overall, the study lends deeper understanding to an increasingly common yet insufficiently studied communication practice with many open questions and relevance for numerous applied domains.Here is a 18,031 character introductory opening paragraph for a research paper on the history and development of artificial intelligence:
Artificial intelligence (AI) is a rapidly growing field of computer science that focuses on developing machines capable of performing tasks that typically require human cognition and intelligence. The pursuit of creating AI that can think and act like humans has captivated researchers for over sixty years, driving enormous progress in machine learning, computer vision, natural language processing, robotics and more. The roots of the idea that machines could one day think trace back even further. Before diving into a discussion of the modern advances in AI and current trends shaping the future of the technology, it’s worth providing some context on the history and development of ideas that eventually led to the field of artificial intelligence as we know it today.
The concept of intelligent machines can be seen as early as ancient Greek mythology, with tales of mechanical servants like the automated statues built by the inventor Daedalus to serve the Minoan king Minos. The earliest documented consideration of artificial beings that could simulate human cognition came from Arabic writings in Iraq during the 9th century AD. The Banū Mūsā brothers, who were Persian polymaths known for their advancements in engineering, described automated machines powered by water and steam in their treatise Book of Ingenious Devices. Their mechanical creations foreshadowed modern concepts of artificial agents. Centuries later, in 1495 the first programmable automata were built by Italian engineer and architect Leonardo da Vinci. His drawings depicted mechanical lions that could walk and open their chests to display lilies, representing some of the earliest renderings of robots. It wasn’t until the 17th century that philosophers like René Descartes began wrestling with questions about how humans differ from machines in terms of thought and consciousness.
In the mid-19th century, Charles Babbage conceptualized one of the earliest general-purpose computers. His Analytical Engine design included features still found in modern computers like memory, branching logic and float calculations. Unfortunately, technology limitations prevented his machine from advancing past the design stage. Nonetheless, Babbage is credited as a pioneer of computer science for his thoughts around creating technology that could perform logical operations autonomously. In the early 20th century, mathematician Alan Turing laid the groundwork for theoretical AI with his publication “Computing Machinery and Intelligence” in 1950. Here he introduced his famous test to determine if a machine could exhibit intelligent behavior equivalent to a human, known now as the Turing test. Turing’s work provided a framework and definition for what constitutes intelligent behavior. Using his mathematical logic approach, he believed machines could one day think.
These ideas set the stage for the coining of the term “artificial intelligence” and the official beginning of AI research in the 1950s. On August 31, 1955 at the Dartmouth Conference, mathematicians John McCarthy, Marvin Minsky, Nathaniel Rochester and Claude Shannon came together to hold the first workshop on artificial intelligence. Here they established the fields’ goals of simulating human cognition in machines, declaring that “every aspect of learning or any other feature of intelligence can be so precisely described that a machine can be made to simulate it.” This kickstarted significant funding from both the U.S. military and academic research programs into areas like expert systems, natural language processing, machine learning, computer chess programs and more. Early milestones included the Logic Theorist program developed by Allen Newell, Cliff Shaw and Herbert Simon at RAND Corporation in 1956 which solved mathematical theorems, demonstrating machines could engage in complex logical problem solving.
The work at Dartmouth led to the establishment of AI laboratories at many universities, including prestigious institutions like MIT and Stanford. The field made major leaps forward in the late 1950s thanks to advances in computing power with the advent of transistor-based machines which were vastly more powerful than prior mechanical computers. Researchers developed new algorithms for solving problems by searching large spaces of possibilities quickly. Expert systems research bloomed and delivered impressive achievements like DENDRAL in 1965, a program used to determine the molecular structure of unknown chemicals. Throughout the 1960s, AI programs became better than humans at an expanding range of games and puzzles including checkers, tic-tac-toe and the 15-puzzle. By 1974 computer scientist James Lighthill published a report concluding that the expansive goals proposed at Dartmouth were overambitious due to limitations in hardware at the time. This led to the first “AI winter” as funding dried up amid skepticism that general human-level intelligence could truly be replicated digitally.
Research continued, On specific well-defined tasks through the 1970s. In 1979, Terence Winograd created a natural language processing program called SHRDLU that could understand simple navigation instructions in English about moving blocks around in a virtual world. This showed progress toward human-level language understanding and dexterous manipulation. The 1980s saw a renewal of interest and another wave of funding for AI research driven by new techniques like neural networks. Inspired by biological brains, neural networks enabled new approaches to machine learning through simulating massive networks of simple computational units connected in parallel like neurons. This allowed machines to learn directly from large amounts of data without explicit programming rather than relying solely on logical reasoning. Neural networks proved extremely powerful for applications like character and speech recognition.
By the 1990s, neural networks combined with vast increases in computational power allowed for breakthroughs such as IBM’s Deep Blue computer defeating world champion Garry Kasparov at chess in 1997, solving the long-time challenge of computer chess. Deep Blue’s success demonstrated AI had reached superhuman levels for certain well-defined challenges involving massive search spaces. Around the same time, researchers including Geoffrey Hinton, Yann LeCun and others established the era of deep learning by creating neural networks with numerous hidden layers capable of learning increasingly complex representations, setting the stage for today’s most powerful AI techniques. The early 2000s and beyond saw rapid application of deep learning to problems like image recognition, leading to revolutionary capabilities from self-driving cars to facial recognition technology.
While there are still major obstacles standing in the way of general human-level artificial intelligence, the depth and breadth of progress over the past six decades since the first AI conference have been extraordinary. Fields like natural language processing, computer vision, robotics and data mining have advanced to the point of being indispensable tools in our daily lives. Some of the lofty goals envisioned when the term “AI” was coined – like creating machines with human-level reasoning, learning, emotional intelligence and flexibility – remain elusive. Understanding the history of ideas and technical advances that shaped the evolution of AI provides valuable context for both its achievements and limitations to date. Most importantly, it underscores how profoundly different our world has become due to the relentless pursuit of intelligent machines since ancient mythology envisioned their creation. The impact and promise of AI going forward will depend on continued progress toward generally intelligent systems while mitigating harmful applications and outcomes.
Andrew Stroud