見出し画像

論文指導 リサーチプロポーザルに向けて(2)

プロポーザルをChatGPTをアシスタントに執筆をする作業をさらに行った。まずは原稿をよませる。

read text #1:

read text #2:

これで大体1800ワードの文章をよみこませた。まあやりとりをするには十分な量である。で

based on text #1 and text #2, write an essay.
とお願いした。エッセイ書いてね、というと普通に書いてくる。

返事は省略。で普通に結論まで書いてくる。これから論文に仕上げるわけであるが、

what is the claim?

すると、次のように返してくる。

There are multiple claims made across both texts #1 and #2, as they cover a range of topics. Here are some potential claims that can be extracted from each text:

つまり、文章の中にclaimになる可能性のあるトッピックがたくさんあって、そこからいろいろClaimがつくれるが?

と言ってくる。このあたりどうだろうか。いわゆるEnglish 101の授業だと、いろんなことが言いたいというのは良くないことで言いたいことを絞りなさい、と教師は指導する。だがいろんな事を 書いたって、最初はいいではないか?というのが前回説明したfree writing   の最初の教えだ。いろんなことを書いた文章をもとにエッセイを書いてくれとたのむと、トッピックが多すぎてclaim が作れない、といってくる。これに対応をしてもらおう。

what is a research question?

一般的にリサーチクエスチョンとは何かを聞いておく。こうすることで答えの精度があがっていく。

A research question is a clear and concise inquiry that a researcher aims to answer through their research. It is typically a specific question that guides the research process and helps the researcher to focus on a particular topic or issue. The research question should be well-defined, relevant, and feasible to answer through research methods. It is a fundamental component of a research study, as it shapes the design, methodology, and analysis of the research. The research question should align with the research aims and objectives and contribute to advancing knowledge in the field.

日本語にしておく。

リサーチクエスチョンは、研究者が自分の研究を通じて解答を求めることを目指す、明確で簡潔な問いです。通常、特定の質問であり、研究プロセスを指導し、研究者が特定のトピックや問題に焦点を合わせるのに役立ちます。リサーチクエスチョンは、明確で関連性があり、研究方法を通じて回答できる範囲で実現可能であるべきです。リサーチクエスチョンは、研究の設計、方法論、および分析を形作るため、研究研究の基本的な構成要素です。研究の問いは、研究目的と目標に合致し、分野の知識を発展させるために貢献すべきです。

というわけでリサーチクエスチョンをすることが非常に大事なことがわかる。では先ほど書いてもらったエッセイにリサーチクエスチョンを問うことはできるのか?

what are the research questions?

すると次のように答えが返ってきた。

There are no explicitly stated research questions in either of the texts. However, there are some implied research questions that could be inferred based on the content of the texts:

いまのエッセイではリサーチクエスチョンを明確に立てられないといってきた。だがこれでいいのである。この感じでの執筆者とChatGPTとの対話の作り方を博士論文プロポーザルを書く人は覚えてほしい。自分は何をして、ChatGPT は何をしてくれるかだ。

次の文章が大事だ。

could be inferred based on the content of the texts:

つまりChatGPTがinfer してくれる、という。infer とは推測をする、という意味で、ChatGPTに聞いてみた。

"推測する"は、証拠と推理に基づいて情報を演繹または結論付けるプロセスを指す動詞です。何かを推測するとき、直接述べられていなくても、利用可能な情報に基づいて結論を引き出したり、論理的な判断を下したりします。この精神的プロセスは、批判的思考や理解の重要な部分であり、明示されたものを超えて状況を理解し、関連性を見出し、解釈することができます。

とこたえてきた。で、claim とresearch question   は何かを聞く。

What is a claim on the essay?

The claim of the essay is   以下省略。

what are the research questions?

There are no explicit research questions stated in the essay. However, some potential implicit research questions that could be inferred from the text are:

  1. What is 省略

  2. How can  省略

  3. How can 省略

  4. What are the key elements of 省略

  5. What is the role of 省略

  6. How can 省略

答えは省略したが、How とWhatの質問を書いているエッセイに問いかけていく。

そして、ここで明らかにしたclaim とresearch questionsを考えて、すでに読んだエッセイをもとに新しいエッセイを書いてくれ、と頼む。

then write a new essay including the claim and research questions based on the essay you read

すると、次のように書いてくる。
The use of technology in education has significantly impacted

しかし、教育におけるテクノロジーについて書いていたわけではない。

my topic is not technology in education, but(以下省略)

といれると、こちらが指定していたテーマに変えて、論文を書き上げる。
答えは省略。

だが、途中で話はスタートアップになっているが、研究対象は大組織だ。そこでもう一度修正を依頼する。

My interest is not in startups, but about rebranding in large incumbent corporations. Please rewrite based on this.

さて、 大体できたので、次の質問をする。

write the rationale for this paper

rationale は論文執筆における専門用語である。ChatGPTに意味を聞いてみよう。

what is rationale in the academic paper?

In an academic paper, the rationale refers to the underlying reasoning or justification for conducting a particular study or research project. It provides a clear explanation of why the research is necessary, important, or relevant, and outlines the context, background, and theoretical basis for the study. The rationale helps readers understand the significance of the research question and the potential contributions the study can make to the existing body of knowledge. It also demonstrates the researcher's understanding of the field and the gaps or issues that the research aims to address. Writing a strong rationale is an essential part of an academic paper, as it sets the foundation for the research and establishes the need for the study.

translate into Japanese
学術論文において、根拠(rationale)とは、特定の研究や研究プロジェクトを実施するための基本的な理由や正当性を指します。研究がなぜ必要で重要で関連性があるのかを明確に説明し、研究の文脈、背景、および理論的基盤を概説します。根拠は、読者が研究の問いの重要性と研究が既存の知識体系に対して提供できる潜在的な貢献を理解するのに役立ちます。また、研究者が分野に対する理解と、研究が対処しようとするギャップや問題を示します。強力な根拠を書くことは、研究の基盤を築き、研究の必要性を確立するため、学術論文の重要な部分です。

ということだ。なぜ(Why)この研究をするのですか?という問いに、なぜなら、と答えていく感じである。

ChatGPTの答えは

The rationale for this paper is to explore the critical role of (省略)であるが、300語ほどでしっかりとこの論文のrationale を含んだ文章を書いてくる。

で大体中身と主張とリサーチクエスチョンとrationale がそろったので、会社の研究所の紀要に論文として提出したいのだが、と相談する。

I need to write an essay on this for the research institute in my company. Please write a long essay based on the above.

すると、論文を書いてくる。

Introduction:
In today's highly competitive business environment, 以下省略。

大体600語くらいのエッセイになる。だが、もう少し展開できそうなので注文をだす。

My concern is that  省略。
Can you please rewrite based on this information?

すると、次のように答える。 おまかせ、という感じだ。

Certainly, here's a revised essay that focuses on the importance of 以下省略。ここまでで、語数は475単語。これをブラッシュアップしていきたい。

please elaborate and specify the essay you just wrote.
elaborateとspecify は論文添削のときの専門用語で、指導教員は赤鉛筆で博士課程の学生の原稿に手を入れるときに使う。elaborate は話の筋がとおっていないときに指導する言葉。specify は主張に対して事例が不足しているときに指導する言葉である。

Certainly. The essay I just wrote focuses on the role of
 やってみたがあまり文字数は増えなかった。理屈ではこのあたりが限界か。
次にアプローチをかえて、論文にデータを加えていくことにした。

explain the primary data and the secondary data in academic research.

論文には二種類のデータがある。これを知らないと、論文は書けない。これについてもChatGPTに聞いてみた。

In academic research, primary data and secondary data are two types of data that researchers use to answer research questions and test hypotheses.
Primary data refers to data that researchers collect directly from their research participants or sources, using methods such as surveys, interviews, focus groups, experiments, or observations. The primary data is original and has not been analyzed or interpreted before. The researcher has control over the data collection process, which enables them to tailor the questions and methods to their research questions and goals.
The advantages of using primary data include that it is specific to the researcher's research question, more accurate, and up-to-date. Moreover, primary data is unique and has not been analyzed or interpreted before, which reduces the likelihood of biased or flawed data.

On the other hand, secondary data refers to data that has already been collected and processed for other purposes. Researchers use secondary data to support or complement their primary data findings, or in cases where primary data collection is not feasible or practical. Sources of secondary data can include academic articles, books, reports, government publications, statistical databases, and historical records.
The advantages of using secondary data include lower costs, easier access, and time savings, as the data has already been collected and organized by others. Additionally, secondary data can provide a broader context for the research topic, offering a valuable comparison point for primary data findings. However, the limitations of secondary data include potential issues with relevance, quality, and accuracy, since the data was not collected specifically for the researcher's study.
In conclusion, both primary and secondary data play crucial roles in academic research. While primary data offers unique, original, and tailored information for the research question, secondary data provides valuable context, support, and comparisons. Researchers often use a combination of both types of data to strengthen their studies and gain a more comprehensive understanding of their research topic. Understanding the differences between primary and secondary data is essential for conducting robust academic research and writing a strong research paper.

 翻訳してみた。

学術研究において、一次データと二次データは、研究者が研究の問いに答えたり、仮説を検証するために使用する2種類のデータです。一次データは、調査、インタビュー、フォーカスグループ、実験、または観察などの方法を使用して、研究者が研究参加者や情報源から直接収集するデータを指します。一次データは独自のものであり、以前に分析や解釈が行われていません。研究者はデータ収集プロセスを管理することができるため、研究の問いや目標に合わせて質問や方法をカスタマイズできます。一次データを使用する利点には、研究者の研究の問いに特化していること、より正確であること、最新の情報であることが含まれます。さらに、一次データは独自のものであり、以前に分析や解釈が行われていないため、偏ったり欠陥のあるデータの可能性を減らします。
一方、二次データとは、すでに他の目的で収集および処理されたデータを指します。研究者は、二次データを使用して、一次データの調査結果をサポートまたは補完したり、一次データの収集が実現不可能または現実的でない場合に使用します。二次データの情報源には、学術論文、書籍、報告書、政府出版物、統計データベース、歴史記録などが含まれます。二次データを使用する利点には、他の人がすでにデータを収集および整理しているため、コストが低く、アクセスしやすく、時間を節約できることが挙げられます。さらに、二次データは、研究トピックの広い文脈を提供し、一次データの調査結果と比較するための貴重な比較ポイントを提供することができます。ただし、二次データの制約には、データが研究者の研究のために特別に収集されていないため、関連性、品質、および正確さに関連する潜在的な問題が含まれます。結論として、一次データと二次データは、どちらも学術研究において重要な役割を果たします。一次データは、研究の問いに対する独自の、オリジナルな、カスタマイズされた情報を提供する一方で、二次データは貴重な文脈、サポート、および比較を提供します。研究者は、両方のタイプのデータを組み合わせて使用することがよくあり、研究トピックについてより包括的な理解を得るために研究を強化します。一次データと二次データの違いを理解することは、堅実な学術研究を行い、強力な研究論文を執筆するために不可欠です。

 とある。現状、ここで指導している研究はかなりのプライマリーデータがあるのだが、その整理はこれからである。それまでにセカンダリーデータをまとめてリサーチレビューを書くことにして、次のような質問をした。

please find the references and put them into the adequate places in the text and create a bibliography of the reference and write it after conclusion.

つまり、いままでに書いたエッセイのテキストに注をつけて、書籍名をレファレンスとしていれて、最後に文献目録をつくってくれ、と聞いている。これでセカンダリーデータをそろえる準備を始めることができる。

ChatGPTは自分で書いた論文に注をつけ始める。だがもう一息なけっかしかでてこないので、次のようなプロンプトエンジニアリングをおこなった。

explain the meaning of specification in the context of academic writing.

ChatGPTは専門用語の定義の質問をすると答えてくれるがそれだけではなくてその答えをコンテキストとして次の質問にも答えるようになる。

In the context of academic writing, specification refers to providing detailed and specific information about a particular concept, theory, method, or data. It involves defining and describing the key features, characteristics, or elements of the subject matter under study. Specifications can include information such as definitions, measurements, procedures, and limitations.
For example, in a research paper on a particular scientific experiment, the specification section would provide details about the materials used, the methods employed, the measurements taken, and the limitations of the study. This information helps readers to understand the study's scope and reliability, and it also allows other researchers to replicate the study and build upon its findings.
Overall, specifications play a critical role in academic writing, as they help to ensure that the information presented is accurate, comprehensive, and replicable. By providing clear and specific details about the subject matter, specifications enable readers to understand the research methodology and results, and they also help to establish the credibility and validity of the study.

翻訳しておく。
学術文書の文脈において、「 specification」は、特定の概念、理論、方法、またはデータに関する詳細で具体的な情報を提供することを指します。これには、研究対象の主要な特徴、特性、または要素を定義し、説明することが含まれます。 specificationには、定義、測定、手順、制限などの情報が含まれることがあります。

例えば、特定の科学実験に関する研究論文では、 specificationセクションでは、使用された材料、採用された方法、行われた測定、および研究の制限に関する詳細が提供されます。この情報は、読者が研究の範囲と信頼性を理解するのに役立ち、また他の研究者が研究を再現し、その結果に基づいて構築することを可能にします。

全体として、 specificationは学術文書において重要な役割を果たします。これは、提示された情報が正確で包括的で再現可能であることを確保するのに役立ちます。対象物に関する明確で具体的な詳細を提供することで、 specificationは、読者が研究方法と結果を理解することを可能にし、また研究の信頼性と妥当性を確立するのに役立ちます

とある。でいまChatGPTで書いているエッセイにおいてspecification に欠けているところを聞いてみた。
can you find any sentence which requires more specification?

In the essay I wrote earlier, a sentence that could benefit from more specification is:

This sentence could be more specific in explaining 省略
ときて修正をしてくる。
Some possible ways to specify this sentence could be:

  • "Furthermore,

  • "Furthermore,

  • "Furthermore,

とspecification が必要なところを指摘してきて、次のように言う。

By providing more specific examples or details, the sentence can become more precise and informative.

より具体的な例や詳細を提供することで、文はより正確で情報性が高くなります。

 このやりとりをして次のように聞く。

will you make a list of references?

するとうれしそうに、次のように答えて、参照文献のリストをはじめる。

Sure, here is a list of references used in the essay:
References:

全部で30冊くらいの文献をあげてきた。この研究は大組織の問題も論じているので、センゲの研究もだが、それに影響を与えたDavid Bohmについてもきいてみた。

do you know David Bohm?

すると答えてくる。

Yes, David Bohm was an American theoretical physicist who made important contributions in the fields of quantum mechanics and philosophy of science. He is particularly known for his work on the interpretation of quantum mechanics and his concept of "implicate order." He also made contributions to the study of plasma physics and the theory of relativity. Bohm is regarded as one of the most important theoretical physicists of the 20th century.

とまあ一般的な話を返してくるが、これがコンテキストとなるので、次の質問をする。

do you know his book wholeness and the implicate order?

Yes, David Bohm's book "Wholeness and the Implicate Order" is a well-known work in the field of quantum physics and philosophy.

ボームの理論はセンゲに受け継がれて、MITで大組織マネージメントの方法としてかなり普及した。この切り口を持ってこれないか聞いてみた。

I want to bring the discussion about Bohm's "Wholeness and the Implicate Order" into the above essay. What do you think?

Incorporating Bohm's "Wholeness and the Implicate Order" into the essay can add a new dimension to the discussion of 省略。

In his book, Bohm argues that reality is not a collection of separate objects but an interconnected whole, where everything is fundamentally related to everything else. This interconnectedness can be seen in the implicate order, which refers to the hidden order of the universe that underlies the manifest world of our sensory experience.
In the context of 省略, this idea of interconnectedness can be applied to the various elements of a 省略。
Furthermore, Bohm's ideas on the implicate order can also highlight the importance of considering the larger context of a 省略.
In conclusion, incorporating Bohm's "Wholeness and the Implicate Order" into the discussion of 省略.

というわけで、ボームの大組織マネージメント議論をいれて論文を書き直してもらうことにした。

combine and rewrite this essay and the previous essay you wrote about Bohm :

エッセイは省略するが、およそ900語になった。これにプライマリーデータの検討をくわえていきたい。どうするかChatGPTに聞いてみた。

what kind of primary data do you recommend?

The type of primary data you collect for your research will depend on your research question and the research design you choose. Generally, primary data can be collected through various methods such as interviews, surveys, observations, experiments, and focus groups. Each method has its strengths and limitations, so it's essential to choose the method that is best suited to your research question and objectives.
If your research question is focused on understanding 以下略。

プライマリーデータの収集に関してあまり面白い案はでてこない。ポイントはリサーチクエスチョンにプライマリーデータで答えることが必要で、そのために方法論を示す必要がある、ということである。

そこで、このエッセイの著者が自分のバックグラウンドをChatGPTに説明をした。

すると500語ほどで、どのようにプライマリーデータを使ってそれを使うのかを説明してきた。省略するが、なかなかの内容だったのでそれを論文のなかに組み込むことにした。次はセカンダリーデータをそろえる。次の質問をした。

will you make a list of references?

Certainly! Here is a list of references for the essay:
References:
として20冊ほどあげてきた。(省略)

最新のものはないか、と聞いてみた。

how about after 2020?
すると、次のように答えてきた。
I'm sorry, I cannot provide a list of references after my knowledge cutoff of 2021. My training only includes information up to that point, and I am not aware of any new publications or developments in the field beyond that date.

とはいうものの、ベストを尽くす、ということでさらに
ChatGPTに聞いてみた。

will you make a bibliography based on the lists you gave and write it according to the alphabetical order?

すると文献目録をアルファベット順になれべてくれた。

Certainly! Here is the bibliography, arranged in alphabetical order:

Bibliography
以下省略。

でここまでの議論をふまえて、論文を参考文献をととのえて書き直してもらった。
can you rewrite your essay by putting the reference into the adequate place in the text, and write a bibliography and place it after the conclusion?

Sure, I can do that for you. Here is the rewritten essay with references and a bibliography:
ということで論文を書き上げた。およそ1000語になった。

さて、これから先は具体的な論文として仕上げる作業にむかうので、どこか関係する学会誌にむけて書き始めたいので、応募できる雑誌を聞いてみた。

please make a list of academic journals that might be relevant.

Sure, here are some academic journals
とこたえて、いくつもの学会誌を紹介してくれた。

さて、だいたいこれでドラフトがそろってきたので、今回の目標の博士論文プロポーザルについて聞いてみた。
how many words the PhD proposal request?


There is no strict word count for a PhD proposal, as the length can vary depending on the specific requirements of the institution or program you are applying to. However, most PhD proposals range from 1,500 to 3,000 words.

データの詳細な検討とリテラチャーレビューがなくて1000語弱なので、量としてはこれを 方法論やプライマリーデータの説明で拡大すればプロポーザルになる。

It's important to follow the guidelines provided by your target institution or consult with your potential supervisor for specific requirements.
A well-structured PhD proposal should typically include the following elements:
良いプロポーザルは次の条件を満たす必要がある。

a clear research question or problem statement, 
ここは丁寧に行ったので大丈夫だろう。

the rationale for the study,
ここも大丈夫だ。

a literature review,
これがこれからの課題。

the proposed research methodology,
これもプライマリーデータをどう集めて分析するかがこれから。

a timeline for the research, and
これはたいした問題じゃない

a preliminary bibliography.
これも着々と進んでいる。

以下、ChatGPTからの戒め。
訳しておこう。

Remember to write concisely, focusing on the key elements of your research, and demonstrating your understanding of the field and the potential contribution your study can make.

研究の主要な要素に焦点を当て、分野への理解と研究ができる潜在的な貢献を示すように、簡潔に書くことを忘れないでください。

 というわけだ。研究したいことがあって、それについてだらだらと書くことができれば、ここで展開したプロンプトエンジニアリングに従えば博士論文プロポーザルに向かうことができる。アイデアや思いつきをどんどん free writing  で書けば、本格的な研究をおこなうための入り口に立つことが出来る。そのあたりを自由に書いてみた。是非自分が自由にかいたエッセイをたっぷりとChatGPTに読んでもらって、ここまでプロンプトエンジニアリングの実例を示したようなやりかたで質疑応答をしながら博士論文プロポーザルのドラフトを仕上げてもらいたい。

この記事が気に入ったらサポートをしてみませんか?