The Artificial Intelligence tools (like ChatGPT, Perplexity, Gemini, Meta, CoPilot and Claude) can provide much useful and nearly instant information about subjects in a concise format, especially when they have been pre-trained on a large volume of information about the subjects.
I wondered if the FREE AI tools could provide me with information about creating a genealogical proof argument that meets the Genealogical Proof Standard.
1) My first Prompt was: "You are a professional genealogist experienced with the Genealogical Proof Standard (GPS). What process and evidence comprises a proof argument that meets the Genealogical Proof Standard?"
ChatGPT4 provided this response:
A proof argument that meets the Genealogical Proof Standard (GPS) must involve a structured process and a combination of evidence that collectively addresses all aspects of a genealogical question. Here are the key steps and components needed to create a GPS-compliant proof argument:
1. Conduct a Reasonably Exhaustive Search
- Process: Gather as much information as possible, including all relevant records, to ensure that the research is thorough and no significant sources are overlooked. This may involve a variety of sources: vital records, census records, wills, deeds, court documents, church records, and other historical documents.
- Evidence: Collect direct, indirect, and negative evidence from diverse and relevant sources that pertain to the question at hand. A broad array of evidence reduces the risk of missing contradictory or corroborative details.
2. Ensure Complete and Accurate Source Citations
- Process: Document each source used in the research process with precise citations, following a standard citation format. Each citation should allow another researcher to locate the same source independently and verify its contents.
- Evidence: Citations should detail the author, title, publication information, location, and specific page or record number within each source. Digital sources should include a permanent URL or repository information.
3. Analyze and Correlate the Evidence
- Process: Carefully examine each piece of evidence in context, understanding its reliability, credibility, and relevance. This includes evaluating the source’s original purpose, assessing its accuracy, and noting any biases or limitations.
- Correlation: Compare and contrast evidence from multiple sources, identifying patterns, contradictions, and supportive details that either strengthen or weaken the argument. Consider factors like proximity to the event, primary vs. secondary nature, and internal consistency.
4. Resolve Conflicting Evidence
- Process: Identify and address any contradictions in the evidence, analyzing the sources’ reliability and the probable cause of discrepancies. This may involve weighing evidence based on the quality of the source or the strength of each piece within the body of evidence.
- Resolution: Provide a reasoned explanation for any conflicts, demonstrating why one interpretation of evidence is more likely correct than others. Without this step, any unresolved conflict weakens the proof.
5. Write a Clear and Convincing Conclusion
- Process: Assemble the findings into a clear, logically reasoned argument. The conclusion should explain the basis of the final determination, detailing how each piece of evidence contributes to answering the research question.
- Argument: Present a coherent narrative or report that leads the reader through the steps of research, analysis, and reasoning used to reach the conclusion. Avoid assumptions not supported by evidence and address any remaining limitations of the argument.
Summary: Building a GPS-Compliant Proof Argument
In sum, a GPS-compliant proof argument requires thorough, well-documented research, critical analysis of all relevant sources, and a logical, evidence-supported resolution. By following these steps, a genealogist can create a defensible and transparent argument that meets the rigorous standards of genealogical proof.
2) My second prompt: "can you provide an example of a successful proof argument using a fictional person, say Frank Seaver (1852-1922), born and lived in Massachusetts, who married, had three sons, and was a combmaker?"
The ChatGPT4 response was:
Certainly! Here’s a fictional example of a successful proof argument for Frank Seaver (1852–1922) that follows the Genealogical Proof Standard (GPS).
Research Question
Who were the parents of Frank Seaver, born in 1852 in Massachusetts, who was a combmaker and died in 1922?
Summary of Known Facts
- Name: Frank Seaver
- Birth Year: 1852
- Birth Place: Massachusetts
- Occupation: Combmaker
- Death Year: 1922
- Residence: Massachusetts
- Family: Married with three sons
Step 1: Conducting a Reasonably Exhaustive Search
To identify Frank Seaver’s parents, I searched multiple sources that could include his parents' names. The research plan included:
- Vital Records: Birth, marriage, and death records in Massachusetts.
- Census Records: U.S. Federal Census records from 1850 to 1920 to track Frank through his life and potentially identify family members.
- Local Histories: Histories of Massachusetts towns known for comb-making, looking for references to Frank and potentially his family.
- Occupational Records: Records of the comb-making industry, as Frank's occupation was specific and could connect him to family members in the same trade.
- Probate and Land Records: Records in Massachusetts for property ownership or inheritance, which might link Frank to family members.
Step 2: Source Citations
Each piece of evidence is fully cited here to ensure traceability:
- 1852 Birth Record: Massachusetts Vital Records, Births, 1852, “Frank Seaver,” Middlesex County; Massachusetts State Archives.
- Marriage Record: Massachusetts Vital Records, Marriages, 1876, “Frank Seaver and Mary Ellis,” Middlesex County; Massachusetts State Archives.
- Death Record: Massachusetts Vital Records, Deaths, 1922, “Frank Seaver,” Middlesex County; Massachusetts State Archives.
- Federal Census Records: U.S. Federal Census, 1860–1920, Middlesex County, Massachusetts.
- Probate Records: Middlesex County Probate Court, Will and Inventory of Nathaniel Seaver (potential father), 1875.
- Land Deed: Middlesex County Deeds, 1870, property transfer from Nathaniel Seaver to Frank Seaver; Middlesex County Registry of Deeds.
Step 3: Analysis and Correlation of Evidence
- Birth Record: The 1852 birth record in Middlesex County lists Frank Seaver, born to Nathaniel Seaver and Rebecca (Jones) Seaver. His birth year and county match the known information about Frank.
- Marriage Record: Frank Seaver’s marriage record from 1876 lists his parents as Nathaniel Seaver and Rebecca Jones, aligning with the birth record.
- Death Record: The 1922 death record for Frank Seaver includes his occupation (combmaker) and spouse (Mary Ellis), confirming the identity match. His parents are also listed as Nathaniel Seaver and Rebecca Jones.
- Census Records:
- In the 1860 census, Frank Seaver, age 8, appears in Middlesex County with parents Nathaniel and Rebecca Seaver, confirming the family grouping.
- In the 1870 census, Frank, age 18, lives in the same county, working as an apprentice combmaker.
- Subsequent census records (1880–1920) show Frank as a combmaker, married to Mary, with three sons, confirming continuity of identity and occupation.
- Probate Record: Nathaniel Seaver’s 1875 probate file names “Frank Seaver, my son,” providing further direct evidence of the relationship.
- Land Record: A deed from 1870 shows Nathaniel transferring property to Frank, suggesting a familial relationship and possible early support in establishing Frank’s career.
Step 4: Resolving Conflicting Evidence
In this case, all records consistently point to Nathaniel and Rebecca (Jones) Seaver as Frank’s parents. No conflicting evidence appears in the search—no other Frank Seaver with a different parentage, birth location, or birth date was found in Massachusetts records during this period.
Step 5: Conclusion
The combined evidence supports that Frank Seaver, born in 1852 in Middlesex County, Massachusetts, who died in 1922, was indeed the son of Nathaniel Seaver and Rebecca Jones. This conclusion rests on:
- Corroborating information from Frank’s birth, marriage, and death records.
- Consistent census records that trace him through life stages and confirm his occupation.
- Direct statements from Nathaniel Seaver’s probate and land records linking Frank as a son.
All sources were analyzed for reliability, correlation, and consistency, satisfying the Genealogical Proof Standard. This proof argument firmly establishes Frank Seaver’s parentage through a well-supported, documented, and reasoned conclusion.
3) My third prompt was: "What information is most important in selecting a research question?"
The ChatGPT4 response was:
Selecting an effective research question in genealogy is crucial to achieving a focused and productive investigation. The following key considerations will help you identify the most important aspects of choosing a research question:
1. Specificity
- A good research question is clear and specific, targeting a single fact or relationship.
- Avoid overly broad questions like “Who were all of my ancestors in the 1800s?” Instead, ask, “Who were the parents of Henry Carringer, born in 1800 in Pennsylvania?”
2. Relevance
- Choose a question that aligns with your broader research goals, such as tracing a direct line, uncovering unknown ancestors, or solving a family mystery.
- Example: If your goal is to verify eligibility for a lineage society, focus on proving a specific ancestor’s parentage or military service.
3. Time and Place
- Incorporate known details about the time period and location relevant to the individual or event.
- Example: “Where was Sarah Feather born in 1804, and who were her parents?” reflects her birth year and place, guiding the search to relevant records.
4. Availability of Records
- Consider whether sources exist to answer the question. Research in areas with sparse records may require broadening the scope or refining the question.
- Example: If local records for a specific town are missing, expand the geographic scope to neighboring counties or regions.
5. Complexity
- Balance the complexity of the question with the time, resources, and skills available. Simple questions (e.g., finding a marriage date) are easier to address than complex ones (e.g., resolving conflicting evidence of parentage).
- Example: Beginners might focus on confirming vital records, while experienced genealogists may tackle multi-generational puzzles.
6. Known Information
- Use what you already know as a foundation. Formulate the question to fill gaps in your research or clarify ambiguous details.
- Example: If you know a couple’s children but not their marriage date, ask, “When and where did Henry Carringer and Sarah Feather marry?”
7. Feasibility
- Ensure the question is answerable with existing resources, considering access to online databases, libraries, or archives.
- Example: A question like “What were the names of John Doe’s grandparents?” may require expanding to indirect evidence if no direct records survive.
8. Avoiding Assumptions
- Phrase the question neutrally, avoiding assumptions that might bias your research.
- Example: Instead of “Was Henry Carringer’s father named John?” ask, “Who were the parents of Henry Carringer, born in 1800 in Pennsylvania?”
9. End Goal
- Think about how answering the question will advance your research. Will it:
- Break through a brick wall?
- Connect generations?
- Clarify a conflicting relationship?
- Verify eligibility for a society or legal claim?
By carefully considering these factors, you can frame a research question that is specific, relevant, and actionable, setting the stage for a successful genealogical investigation.
5) I asked the same questions on Perplexity, and received similar answers. I recommend doing several AI tools because they may have been trained on different reference materials.
Perplexity also provides links and summary information from the Internet. In this case, their answers were informed by:
6) Is this information complete and accurate? We always have to confirm information that has been summarized, analyzed and presented by the AI Tools, and they may not be complete or accurate. This information looks pretty good to me, and appears to be derived from resources that experienced genealogists know to use for the GP S and crafting a genealogical proof argument.
I like that the AI tools created a "Research Question" to answer in the second Prompt - I intentionally didn't ask it to do that. Note that I gave the AI tool very limited information. Although the details of the Frank Seaver answer are intentionally wrong, it's an excellent example. The third prompt asked about the best features of a Research Question. I was surprised that ChatGPT4 used names and dates of my own ancestors as the examples in forming a Research Question.
There is one book reference and seven online references, including a YouTube video) in the list of sources from Perplexity. I don't know why it didn't reference Elizabeth Shown Mills book and/or website, Evidence Explained. It may be as simple as it wasn't pre-trained on the book.
==============================================
Copyright (c) 2024, Randall J. Seaver
Please comment on this post on the website by clicking the URL above and then the "Comments" link at the bottom of each post. Share it on Twitter, Facebook, or Pinterest using the icons below. Or contact me by email at randy.seaver@gmail.com. Note that all comments are moderated, so they may not appear immediately.
