Jump to content
Patrick M. Lofft

plural sons and, or, daughters in the census witness sentences

Recommended Posts

Does TMG v7.03 have an internal methodology to recognize plural sons and, or, daughters in the census witness sentences?

 

If so, how do I turn it on?

 

TMG v7.03 lists all the family’s sons or daughters separated by commas in the witness sentence.

 

At this time, I edit the census witness sentence for each son/daughter when the census lists multiple sons and, or, daughters.

 

As the witness sentence comes: [RF:son] was listed as a son in [R:Head of Family]'s household on the 1930 Census <[L]>. <[M3].>

 

As I revise the witness sentence: [RF:son] were listed as sons in [R:Head of Family]'s household on the 1930 Census <[L]>. <[M3].>

 

Alternatively, how do you handle this?

 

Merry Christmas and a healthful New Year.

 

Patrick

Share this post


Link to post
Share on other sites

The variables [R:rolename] and [W], when used in the sentence for that witness, produce only the name of the one person. The other role variables, like [RF:rolename] will produce a list of all the people with that role, as you found.

 

So, the question is, what result do you want? I use the [W] variable, which produces the full name if it's the first sentence in the paragraph, or the pronoun "He" or "She" if not. Often I edit the sentence locally to manually insert "He" or "She." Perhaps a better solution, which I may adopt, is to create both male and female sentences and use "He" and "She" instead of any variable.

 

If you want all the sons listed in the narrative of each one (that seems odd to me) you can do that, but you need two roles: Son for when there is only one, and Sons for when there is more than one. One role would have a sentence for a single son, and the other would have a sentence for mulitple sons. There is no way to automate this with a single role.

Share this post


Link to post
Share on other sites
The variables [R:rolename] and [W], when used in the sentence for that witness, produce only the name of the one person. The other role variables, like [RF:rolename] will produce a list of all the people with that role, as you found.
Terry is absolutely correct with the above statement, but I emboldened an important part of his text in case it was not clear. Thus [R:Son] used in the sentence for the Son role will give only one person. But the same [R:Son] variable used in a sentence for another role (e.g. the sentence for Principal, Witness, Daughter, Head of Family, etc.) will in that case produce a list of all the people with the role of Son. This same variable works differently depending on the sentence in which it is used.

 

Note that the various pronoun versions of role variables do automatically recognize gender and plural.

[RP:role] Pronoun (he/she/they) of role.[RS:role] Possessive Pronoun (his/her/their) of role.[RM:role] Objective Pronoun (him/her/them) of role.

As Terry mentioned, there is nothing else to aid with whenever a role might have multiple people assigned. We might wish for a mechanism for plural roles (maybe like the single vertical bar to indicate when there is a second Principal "[P] ") but no such feature for roles currently exists.

 

If you want to list all the sons in the sentence for the Son role, depending upon how you have defined and used your roles you might be able to use the [WO] (Witness Other) variable. It includes all the linked witnesses regardless of their role but excludes principals and this person if used in the sentence for a witness. Suppose you had the Head of Family as the Principal and all other members of the household linked as witnesses using various roles (Son, Daughter, MothInLaw, Lodger, etc.) In that case your sentence text for the Son role might be something like:

[R:Son] was listed as a son in [R:Head of Family]'s household on the 1930 Census <[L]>. <Other individuals enumerated in that household were [WO].>

Of course this sentence also has a problem if there is only one "Other" witness .

 

If you actually only want the Given name of only this one son to output in the Son sentence you might consider what I do. I use a split witness memo and put the son's name as spelled in this enumeration as the first witness memo part. Then my son role uses the [WM1] variable instead of a [RF:role] construct something like:

[WM1] was listed as a son in [R:Head of Family]'s household on the 1930 Census <[L]>. <[WM2].>

Hope this gives you ideas.

Share this post


Link to post
Share on other sites

I set up two different roles. One for the singular son and one for the plural sons. That way I can make the sentence read correctly for each without having to make a custom sentence.

 

[RF:son] was listed as a son in [R:Head of Family]'s household on the 1930 Census <[L]>. <[M3].>

 

[RF:sons] were listed as sons in [R:Head of Family]'s household on the 1930 Census <[L]>. <[M3].>

 

 

Sheila Altenbernd

Share this post


Link to post
Share on other sites

Another approach, I use Son01, Daughter01, Son02, Daughter02, etc.

 

If the first child is a boy, I use Son01, if it is a girl I use Daughter02. If the second child is a boy, I use Son02, a girl, Daughter02...

 

 

My head of household sentence then says: Also living in the household were: <his wife, [R:wife]><, his son, [R:son01]><, his daughter, [R:daughter01]><, his son, [R:son02]><, his daughter, [R:daughter02]><, his son, [R:son03]><, his daughter, [R:daughter03]><, his son, [R:son04]><, his daughter, [R:daughter04]><, his son, [R:son05]><, his daughter, [R:daughter05]><, his son, [R:son06]><, his daughter, [R:daughter06]><, his son, [R:son07]><, his daughter, [R:daughter07]><, his son, [R:son08]><, his daughter, [R:daughter08]><, his son, [R:son09]><, his daughter, [R:daughter09]><, his son, [R:son10]><, his daughter, [R:daughter10]>

 

That will read "Also living in the household were: his wife, Rebecca Jones, his son, James Jones, his daughter, Samantha Jones, his daughter, Amanda Jones, his daughter Jennie Jones, his son Adam Jones...

 

 

That way the children print in the HOH sentence in the order they are listed in the census.

 

I prefer to only use one person for the individual sentence, so it is "[W] was enumerated...." for all roles other than the HOH.

 

 

Just thought I'd share that, since it might give you more ideas on how to handle it. The more ideas you have upfront, the less cleanup you have later when a better mousetrap comes along... :santa:

Share this post


Link to post
Share on other sites
The variables [R:rolename] and [W], when used in the sentence for that witness, produce only the name of the one person. The other role variables, like [RF:rolename] will produce a list of all the people with that role, as you found.

 

So, the question is, what result do you want? I use the [W] variable, which produces the full name if it's the first sentence in the paragraph, or the pronoun "He" or "She" if not. Often I edit the sentence locally to manually insert "He" or "She." Perhaps a better solution, which I may adopt, is to create both male and female sentences and use "He" and "She" instead of any variable.

 

If you want all the sons listed in the narrative of each one (that seems odd to me) you can do that, but you need two roles: Son for when there is only one, and Sons for when there is more than one. One role would have a sentence for a single son, and the other would have a sentence for mulitple sons. There is no way to automate this with a single role.

Terry,

We've been through some of this before, but I did not realize that [R:rolename] would produce just one name, while other role variables such as [RF:rolename] would produce the list of all who had the specified role.

 

Is there any logic to that difference? It seems fundamentally inconsistent.

 

As others have pointed out, you have to create extra roles, or custom edit the role sentence to handle one or more people in a specific role, if you want the role sentence to read correctly. That looks like a fudge to get around a poor design decision on how to handle role references.

 

It could also require you to retrofit a tag when you decide to add someone in a particular role. For example, you have a census with several "pupils" listed, only one of which you recognize as a relative. So you link that person to the role of "pupil", and make that person "principal" (because that is the only person on the census you are interested in). The role sentence might read "[RF:pupil] was listed as a pupil in the school of [R:headmaster]...". Later you find another relative who was a pupil at that school, so you link that person to the role of "pupil". Now the sentence will come out "Jane and Mary was listed as a pupil ...". Now you need to go back and create a role of "pupils" if you want to get "Jane and Mary were listed as pupils..." or create roles "pupil1", "pupil2", etc., as suggested by several people. (I've had a situation where I found a relative who was a pupil at a boarding school, and later found another relative recorded as a pupil in the same census.)

 

It seems a better solution might be to have a role code that would indicate you wanted to list all individuals in the role sentence; otherwise only the current one being reported would be used.

 

For example, the letter "M" (if "M" isn't already assigned a meaning) could be used to indicate multiple people are to be listed. For example, [RF:rolename] would list only one person (the same as [R:rolename]), while [RFM:rolename] would list all those with that role. (It might also be useful to have a logical variable indicating whether one or more than one person had the role, Which could be used to select whether "was" or "were" would be used in the sentence. Things could get even more complicated if TMG wanted to support languages that had different spellings of the word for "they" and "pupils" depending on whether "they" included boys or just girls.)

 

There might be a better way to solve this problem, but I think the current logic for handling these sentences is not reasonable. If the individual had the role of, for example, "pupil" (that is what the census said), then that is what we should link the person to. The role sentence for that person should not have to change depending on whether someone else is linked to the same role.

 

Pierce

Share this post


Link to post
Share on other sites

Pierce,

 

Would my new wish item here work for your purposes? I share your desire to make it easier to construct sentences based on whether there was one or more person assigned this role. I realize my proposal does not provide a way to ask that only one person be output or ask for all persons. I am also concerned that any proposal does not break current sentence behavior. Maybe my construct could be interpreted as you desire whenever it is used in the one sentence for a person with that same role. For example, if [RF:Son] was used in the Son role sentence for a person assigned the role of son, then might specify to only output this one person, but for backwards compatibility [RF:Son] would give you however many there are, and would only output if there are multiple. And the current behavior of [R:Son] to only give one person in a sentence for a Son role could be enhanced by using to get multiples.

 

Just trying to come up with a clear-cut proposal.

Share this post


Link to post
Share on other sites
... I did not realize that [R:rolename] would produce just one name, while other role variables such as [RF:rolename] would produce the list of all who had the specified role. Is there any logic to that difference? It seems fundamentally inconsistent.
Pierce,

 

As I mentioned above, this variable is consistent with all other role variables in outputting however many people there are if used in any sentence other than the sentence for that same specific role. Only within that one sentence does this one variable work differently.

Share this post


Link to post
Share on other sites
Pierce,

 

Would my new wish item here work for your purposes? I share your desire to make it easier to construct sentences based on whether there was one or more person assigned this role. I realize my proposal does not provide a way to ask that only one person be output or ask for all persons. I am also concerned that any proposal does not break current sentence behavior. Maybe my construct could be interpreted as you desire whenever it is used in the one sentence for a person with that same role. For example, if [RF:Son] was used in the Son role sentence for a person assigned the role of son, then <[RF:Son]&> might specify to only output this one person, but for backwards compatibility [RF:Son] would give you however many there are, and <&[RF:Son]> would only output if there are multiple. And the current behavior of [R:Son] to only give one person in a sentence for a Son role could be enhanced by using <&[R:Son]> to get multiples.

 

Just trying to come up with a clear-cut proposal.

I think this will allow the creation of a sentence template that will work with one or multiple people linked to a role. At the present time, it does not seem possible to create a sentence template that will create a proper English sentence in either case - it seems you have to create additional, redundant roles and you have to know if you have one or more people in a real-world role before assigning the TMG roles (e.g. "pupil" vrs "pupils").

 

I don't think it solves my original problem. If there are several people in a particular role and I want all tags a person is associated with to be printed in a Journal report, the standard role sentences will not come out correctly. I want "John was listed as the son of Sam Smith..." and there does not seem to be a way to do that if there are more than one "son", unless I custom edit the tag sentence or create multiple roles.

 

Ideally I should be able to link each individual to the correct role, and we should be able to use a default role sentence that references (by name or by pronoun) just that individual. It should also be able to create a sentence that uses a list of names or collective pronoun. For my use, I would only use a single name or pronoun for my default sentences for all roles. Other users may want to produce sentences with a list of names or collective pronoun, possibly using a sentence template like you propose.

 

Pierce

Share this post


Link to post
Share on other sites
I don't think it solves my original problem. If there are several people in a particular role and I want all tags a person is associated with to be printed in a Journal report, the standard role sentences will not come out correctly. I want "John was listed as the son of Sam Smith..." and there does not seem to be a way to do that if there are more than one "son", unless I custom edit the tag sentence or create multiple roles.

 

Ideally I should be able to link each individual to the correct role, and we should be able to use a default role sentence that references (by name or by pronoun) just that individual.

Pierce,

 

You can do that now, but only if you are willing to accept the full name, not just the first name, when a name appears. Use either the [W] or [R:rolename] variable.

 

I'd like to see the ability to do the same with the given name forms too, and that is on the wishlist.

Share this post


Link to post
Share on other sites
Would my new wish item here work for your purposes?
... I don't think it solves my original problem. If there are several people in a particular role and I want all tags a person is associated with to be printed in a Journal report, the standard role sentences will not come out correctly. I want "John was listed as the son of Sam Smith..." and there does not seem to be a way to do that if there are more than one "son", unless I custom edit the tag sentence or create multiple roles.
Pierce,

 

You are correct that the "standard" role sentences will not come out as you wish, unless the standard sentence uses [R:role] which does work in the role sentence exactly equivalent to the [P] variable in standard Principal sentences. However, as the capability for roles was added after a lot of the standard sentences were created, there are not a lot of "standard" sentences with named role variables. Creation of roles usually implies creation of custom sentences.

 

For me, I don't think I use a single "standard" sentence any more, for either roles or Principals. I am more concerned with proposing a new feature that would make it possible to construct "custom" sentences that will do what I want. I believe that my role variables proposal to distinguish one from many being assigned, and this person from the others in that role sentence, would enable constructing much more readable and useful "custom" sentences. For backwards compatibility I don't expect the "standard" sentences or functionality to change.

Share this post


Link to post
Share on other sites
... the standard role sentences will not come out correctly.

What "standard role sentences," Pierce?

 

u are correct that the "standard" role sentences will not come out as you wish, unless the standard sentence uses [R:role] which does work in the role sentence exactly equivalent to the [P] variable in standard Principal sentences.

All the standard role sentence do use the "equivalent to the [P] variable in standard Principal sentences" because they are exactly the same as the standard Principal sentences, including the use of the [P] variable. :)

 

If Pierce has role sentences that work otherwise they were created by the user, or they were imported from UFT.

Share this post


Link to post
Share on other sites
... the standard role sentences will not come out correctly.

What "standard role sentences," Pierce?

 

u are correct that the "standard" role sentences will not come out as you wish, unless the standard sentence uses [R:role] which does work in the role sentence exactly equivalent to the [P] variable in standard Principal sentences.

All the standard role sentence do use the "equivalent to the [P] variable in standard Principal sentences" because they are exactly the same as the standard Principal sentences, including the use of the [P] variable. :)

 

If Pierce has role sentences that work otherwise they were created by the user, or they were imported from UFT.

 

Terry,

You are right: I have lots of role sentences that were imported from UFT. These sentences would work quite well in TMG if TMG were to support "[RF:/rolename]", etc. the same way as "[PF]", etc. A work-around for some of my sentences might be if there were "[WF]", etc. available, but that would not solve all my issues.

 

I still think that the way TMG implements substitution for "[RF:rolename]" is not logical. I do not see how you can create a sentence using this variable that will be correct English when there may be one or more than one person linked to that role. It will only work if you ensure that only one or only more than one person is linked to that role. There are many events where there could be one or many people with the role of "son", "beneficiary", or even "witness" (in the real-world marriage or legal trial sense), etc. and we should not have to have redundant roles defined just to get around an illogical TMG convention.

 

On a related note, I've seen several people comment on how they use "redundant" roles, such as "son1", "son2", etc. so they can create role sentences or memo text that reference several people with the same real-world role, but with different attributes (e.g. age). This technique has the advantage that if you create an index of people for a report, the index will be sure to use consistent names and dates for various references to an individual. (Normal index entries in memo text require the user to be sure that the name and birth/death dates in the index match the "current" values that TMG maintains.)

 

It would be nice if the specification of index entries in memo text could be separated from role references. In UFT, you specify an index entry using the individual number. You can then update the person's name and dates without worrying whether you have index entries in various memo text that have to be updated to be compatible. This would also allow you to enter "incorrect" names in the memo text but have the correct name in the index. For example, a census entry was transcribed as "Francis [sic], Daughter, ..." should have an index entry referencing "Frances". The girl was never called "Francis" by anyone (except the census enumerator who did not read some handwriting correctly), so it is not a valid alternate name. But I don't know if Bob has given this idea further thought since I suggested it to him a couple of years ago.

 

I am certainly getting a better feel for how TMG handles roles, and various techniques for using them. I just hope some of our wish list items on this topic get implemented.

 

Pierce

Share this post


Link to post
Share on other sites
I still think that the way TMG implements substitution for "[RF:rolename]" is not logical. I do not see how you can create a sentence using this variable that will be correct English when there may be one or more than one person linked to that role. It will only work if you ensure that only one or only more than one person is linked to that role.

Pierce,

 

I think the simple answer is that in TMG the [RF:rolename] variable was never intended to be used in the sentence for that role. TMG's standard sentences all use the [P] or [W] variables, which automatically substitute the pronoun "He" or "She" in useful ways. The [R:rolename] variable does the same. But the [Rx:rolename] variations do not. So TMG's design apparently intends that references to the current witness will use either the full name or the pronoun.

 

I don't find adequate - I think there needs to be a way to get just the first name, which, as I said, is on the wishlist.

Share this post


Link to post
Share on other sites
I still think that the way TMG implements substitution for "[RF:rolename]" is not logical. I do not see how you can create a sentence using this variable that will be correct English when there may be one or more than one person linked to that role. It will only work if you ensure that only one or only more than one person is linked to that role.

Pierce,

 

I think the simple answer is that in TMG the [RF:rolename] variable was never intended to be used in the sentence for that role. TMG's standard sentences all use the [P] or [W] variables, which automatically substitute the pronoun "He" or "She" in useful ways. The [R:rolename] variable does the same. But the [Rx:rolename] variations do not. So TMG's design apparently intends that references to the current witness will use either the full name or the pronoun.

 

I don't find adequate - I think there needs to be a way to get just the first name, which, as I said, is on the wishlist.

Terry,

Aside from not being adequate, I think the way the [RF:rolename] and similar role codes are implemented are fundamentally wrong from an English grammar point of view (and is even more wrong for the grammars of other languages).

 

You cannot have [RF:rolename] do something if you might have one or you might have more than one person linked to "rolename". If your sentence template was "[RF:rolename] was ...", then if more than one person is in the role of "rolename", the result would be "John and Jim was ...". For the "Principal" roles (not really proper terminology), you can have either just one person, represented by "P" or two people, represented by "P" and "PO" codes. You can therefore test for the existence of "PO" to decide if the sentence will use a singular or a plural version of verbs. I don't think there is a way for the sentence to detect if there is just one or more than one person linked to regular roles. At least one proposal has been offered to add some syntax to do that but that may not even be on the wish list yet.

 

Pierce

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×