Organizing papers
Like writing in any other genre, a paper in analytical linguistics has a precise, formalized style that readers expect it to follow. Following the expected formula is an important aid for readers, who are faced with the formidable task of coming to terms with complex reasoning based on unfamiliar data. Understanding the mechanics of a paper that does this successfully is essential to good academic writing. Most crucially, authors need to keep in mind that a paper is an argument constructed around a substantive falsifiable thesis, and it must carefully lay out that thesis and present the supporting evidence for it in a logically-sequenced progression of arguments, usually organized into numbered subsections.
This is typically done in a fairly mechanical fashion, beginning with an introduction (discussed in §1 below), moving on to the body of the paper which presents the argumentation (§2), and then summarizing and expanding on the implications of the findings in a conclusion (§3).
1. The introduction
The introduction to a paper serves to orient the reader and to give a clear picture of the context for the research, the central claim being made, and the arguments that are being marshalled in support of the claim. It should serve as a roadmap for the reader: a well-developed set of expectations is essential for navigating through an academic article. Typically, an introduction has three components—a lead-in, a thesis statement, and summary of the argumentation and the contents of the paper.
1.1. The lead-in
The lead-in to the paper should contextualize the research, telling the reader what the general topic area of the paper is and setting up the general problem under discussion. The reader needs to know why the paper is being written and why it is worth reading. For the type of paper published in IJAL, it is often a good idea to start with a bit of data or an illustrative example to give the reader a sense of the issue under discussion.
Authors need to remember that the abstract to the paper can not fill in for the introduction: the abstract is a separate piece of writing and the paper must be able to stand without it.
1.2. The thesis statement
The thesis statement is a clear, concise statement of the analytical or descriptive point that the paper sets out to prove. (The thesis statement for this discussion is in bold in the first paragraph above.) A thesis statement must be a proposition—that is, it must be something that can be proved/disproved by the argumentation in the rest of the paper. Self-referential statements like “this paper is about case-marking” are not going to make good theses for sucessful papers.
1.3. The key points
Once the thesis statement is made clear, the key arguments in support of that thesis should be stated in as concise a form as possible. It isn’t necessary to go into great detail here, but it is necessary for the reader to be able to assess the type of evidence and argumentation being brought to bear on the thesis, and to give the reader a firm idea of what is coming.
1.4. The roadmap
Many writers choose to conclude the introductory section with a paragraph explicitly laying out the contents of the paper to come, section by section. This can, optionally, be combined with setting out the key points in the paper as described in §1.3. As always, the goal here is to give the reader guidance. A good academic article is not like a good mystery novel—there are no unexpected plot twists, and no surprises at the end.
2. The body
The bulk of the paper is, of course, the body. This is where the author sets out the arguments for the thesis in detail, and involves the most complex organization. Typically, the body is divided into sections and subsections, organizing the information the reader needs in the most useful way possible. There are a lot of ways this can go wrong, but the most common error writers (of all levels of experience) make is choosing an order that forces them to present data or arguments that presuppose something that doesn’t come until later in the paper. It is important to remember that a typical reader has little or no pre-existing knowledge of the subject matter or the language. Organizing a paper in a way that doesn’t recognize this, and which fails to present necessary background for the discussion or which doesn’t build from simplest to most complex argumentation is a recipe for failure.
2.1. Linguistic background
Papers in IJAL almost invariably deal with data taken from individual languages that most readers will be unfamiliar with. Understanding these data requires a certain certain general background knowledge of the languages’s structure and typological profile, as well as a specific knowledge of the type of structure being discussed in the paper. Thus, if the paper is about “anomalies” in the use of case-marking to indicate grammatical relations, then the basic morphosyntax of grammatical relations in the language and a sketch of how case is marked will be essential to understanding the ideas that will be presented in the argumentation in the rest of the paper. The traditional place to present these preliminary facts is in the first section of the body of the paper. Authors should also note that IJAL, in addition to being a forum for linguistic research, is also a repository for the presentation and preservation of facts about under-described languages, so this background section can include somewhat more detail about the general grammar of the language than is strictly necessary for the specific purposes of the paper—within reason, of course, and respecting the severe constraints on space that come with paper publication.
2.2. Outline numbering
Linguistics as a genre tends to make heavy use of titled sections and subsections with outline-style numbering. This removes the burden of writing a lot of verbose transitional material to get from one idea to another, but in exchange it requires a lot of careful planning so that the reader can see the connection between sections and subsections clearly as the discussion moves from topic to topic.
The introductory section of the paper can optionally be included in the numbering as section 1, or numbering can begin with the body of the paper.
Probably the most common errors people make in numbering are:
- Leaving a superordinate section empty of text and moving directly to a subsection heading
- Creating a single subsection of a given level under a heading
Both of these practices are bad form and, at best, lead to unnecessary structure in a paper, and, at worst, lead to confusion on the part of the reader.
2.2.1. No empty super-sections
If the argumentation at any point in the paper is complex enough that there is a need to divide it into two or more sub-arguments in separate sub-sections, then it is complex enough for the division to need to be explained, even if that explanation is simply a heads-up to the reader stating explicitly that the divison has been made and listing the coming subsections. Authors should not rely on section headers to do the work for them: section headers are like headlines—terse, elliptical, and sometimes misinterpreted.
2.2.2. No unpaired sub-sections
On the flip side of this, if the argumentation at any point in the papers is not complex enough to divide into two or more subsections, then there is no need for a subsection heading. The topic can be dealt with under the superordinate heading. If the topic really feels like it doesn’t fit without a sub-heading, then it is probably in the wrong place in the paper.
3. Conclusion
The conclusion should do a good job of summarizing the paper, restating the thesis, referring back to key points, summarizing findings, and possibly even repeating key examples. Readers should be able to jump to the conclusion before reading the paper as a whole and get a good sense of what the paper is about and what the original contribution of the paper to the field is.
Over and above providing a summary of the findings, the conclusion can also be a discussion of the implications of these findings. Authors may chose to expand upon some of the key ideas or theoretical implications of the paper, or may outline specific proposals or projects for furthering this line of research. Weak, general comments about the need for “future investigation” or vague references to “remaining problems” without prospects for their solution are a poor way to finish.
4. Good academic writing
The key to good academic writing is to remember that a journal article is an argument built around a single thesis or proposition, and that everything in the paper is devoted to helping the reader understand that proposition and the compelling reasons for believing it to be true. In order to accomplish this successfully, the reader has to have some background about the thesis and to know why it is important (and to whom), and also has to know from the outset how the author is going to proceed and what kinds of evidence will be marshalled in its support.
Once the paper is underway, the reader needs to be given the requisite background knowledge about the language under discussion, and then needs to be walked step by step through the arguments and evidence in support of the thesis. Once this is done, a concluding section recapping the paper will help concretize these points in the reader’s mind. At this point, the author may want to go a little further afield and discuss the bigger-picture implications of the thesis, perhaps with an eye towards underscoring the importance of what has been said and reassuring the reader that their time with the paper has been well-spent.
Of course, we do need to recognize that good academic writing isn’t always as mechanical and formulaic as all this might sound, and experienced writers, as in any other genre, may depart from some of these practices to good effect. Even then, these departures are made in support of the primary goal of scientific writing of any type—the presentation of data and argumentation in support of an original thesis that represents a genuine contribution to knowledge.
Presenting data
Guidelines for interlinear examples are listed here.
When presenting data in blocks or datasets in a linguistics paper (or, for that matter, presenting definitions or lists in inset blocks), it is important to keep in mind two basic principles:
- the data must be presented first, as soon as possible after first mention, before being discussed in detail;
- the data must be introduced in a general way alerting the reader as to why they are there, what they are intended to illustrate, and what the reader should look for in the dataset, and they should be immediately explicated afterwards, highlighting the specific details in the dataset that are of interest, and how these contribute to the paper.
The first principle ensures that readers do not have to flip back and forth between the data and the text, and that they do not have to follow a discussion of something hitherto unseen. The second principle prepares the reader ahead of time for what is coming and makes it easy to find the salient features of a dataset, and afterwards ensures that readers understand the examples and how you are interpreting them for the purposes of your writing.
An example for this type of writing might be drawn from a discussion of causative constructions, where the aim is to illustrate the morphological causative and introduce to the reader to the idea that there are some typological problems with the traditional distinction drawn between the semantic roles of Agent and Causer. We would want to begin this discussion by giving an example of a simple verb from Upper Necaxa Totonac, followed by a sentence containing its causative derivative, as in (1):
| (1) | a. | kit naḭkskúxa čuːwá waːtsá | |||
| kit | na–ḭk–skúx–a | čuːwá | waːtsá | ||
| I | FUT–1SG.SUB–work–IMPF | now | here | ||
| ‘I’ll work here now.’ | |||||
| b. | liːla̰ʔapuːcíː akšní ḭškintamaːskuxúː a̰ːxcananú | ||||
| liː–la̰ʔapuːcíː | akšní | ḭš–kin–ta–maː–skux–úː | a̰ːxcananú | ||
| INST–be.sad | when | PAST–1OBJ–3PL.SUB–CS–work–CS | back.then | ||
| ‘It was sad when they made me work back then.’ | |||||
(1) contrasts the monovalent verb skux- ‘work’ with its causative derivative, maːskuxúː ‘make sby work’, which is bivalent. The new semantic actant in (1b) becomes the subject, and the original subject of the base — generally referred to as the “Causee” — is realized as a direct object. The new semantic actant has the role of Causer, which differs from Agent by virtue of being the initiator of an unspecified event that in turn triggers the event expressed by the verbal base (Langacker 1987). In languages with morphological causatives, verbs whose subjects express prototypical Agents tend to be underived stems whereas those whose subjects are clearly Causers tend to be derived. However, both within and across languages there is a certain fuzziness as to where the line between the two roles is drawn, and which predicates are derived or underived.
How not to present data in a paper
Do not discuss data that you haven’t presented yet
WRONG
Upper Necaxa Totonac displays some apparent irregularities in the expression of subject and object agreement in cases where the subject is first- or second-person and the object is second- or first-person, and one or both is plural, as seen in (2). These forms are three-way ambiguous; the 1 > 2 form in (2a), for instance, which consists of the first-person subject prefix ḭk-, the plural object marker kaː-, and the second-person object suffix -n, has not only the expected 1sg > 2pl reading, it is also used in situations where 1pl > 2pl and 1pl > 2sg. Likewise, the form in (2b) has three readings, 2sg > 1pl, 2pl > 1pl, 2pl > 1sg, although in this case it consists of a non-compositional configuration of affixes, the first-person object prefix kin-, the first-person plural subject suffix -w, and the reciprocal prefix laː-.
| (2) | a. | ḭkaːtṵ́ksnḭ |
| ḭk–kaː–tṵks–nḭ | ||
| 1SG.SUB–PL.OBJ–hit–2OBJ:PFV | ||
| ‘I hit you guys’ or ‘we hit you guys’ or ‘we hit you’ |
| b. | kilaːtṵ́kswḭ | |
| kin–laː–tṵks–wḭ | ||
| INST–RCP–hit–1PL.SUB:PFV | ||
| ‘you hit us’ or ‘you guys hit us’ or ‘you guys hit me’ |
RIGHT
Upper Necaxa Totonac displays some apparent irregularities in the expression of subject and object agreement in cases where the subject is first- or second-person and the object is second- or first-person, and one or both is plural. These forms are three-way ambiguous, as seen in (3), which shows a verb expressing action of the first-person on the second (1 > 2):
| (3) | ḭkaːtṵ́ksnḭ |
| ḭk–kaː–tṵks–nḭ | |
| 1SG.SUB–PL.OBJ–hit–2OBJ:PFV | |
| ‘I hit you guys’ or ‘we hit you guys’ or ‘we hit you’ |
The form in (3) consists of the first-person subject prefix ḭk-, the plural object marker kaː-, and the second-person object suffix -n. This is the form expected for the 1sg > 2pl reading, but it is also used in situations where 1pl > 2pl and 1pl > 2sg.
In a similar vein, the form in (4) expresses action of 2 on 1 and has three readings, 2sg > 1pl, 2pl > 1pl, 2pl > 1sg:
| (4) | kilaːtṵ́kswḭ |
| kin–laː–tṵks–wḭ | |
| INST–RCP–hit–1PL.SUB:PFV | |
| ‘you hit us’ or ‘you guys hit us’ or ‘you guys hit me’ |
Unlike (3), however, this form consists of a non-compositional configuration of affixes—the first-person object prefix kin-, the first-person plural subject suffix -w, and the reciprocal prefix laː-.
Do not concentrate all your data into a single long set then discuss it
WRONG
Upper Necaxa Totonac has a total of six valency-increasing morphemes, two causatives and three applicatives:
| (5) | a. | kimaːɬkaːkníːya̰ | |||
| kin–maː–ɬkaːk–niː–ya̰ | |||||
| 1OBJ–CS–hot–CS–IMPF:2SG.SUB | |||||
| ‘you are making me hot’ | |||||
| b. | ma̰ʔajikwánlḭ tsa̰má ʔawá̰ča̰ | ||||
| ma̰ʔa–jikwán–lḭ | tsa̰má | ʔawá̰ča̰ | |||
| STM–afraid–PFV | that | boy | |||
| ‘he frightened the boy’ | |||||
| c. | na̰kšo̰ʔonḭyáːn ḭštapáɬ | ||||
| na–ḭk–šo̰ʔó–nḭ–yaː–n | ḭš–tapáɬ | ||||
| FUT–1SG.SUB–pay–BEN–IMPF–2OBJ | 3PO–price | ||||
| ‘I’ll pay you its price’ | |||||
| d. | šuːnúːk naliːʔeːchiːkán ša̰ː | ||||
| šuːnúːk | na–liː–ʔeː–chiː–kan | ša̰ː | |||
| bark | FUT–INST–back–tie–IDF | sweatlodge | |||
| ‘they will tie it onto the sweatlodge with bark’ | |||||
| e. | nakinta̰ːpína̰ | ||||
| na–kin–ta̰ː–pin–a̰ | |||||
| FUT–1OBJ–CMT–go:2SUB–IMPF:2SG.SUB | |||||
| ‘you will go with me’ | |||||
| f. | tala̰ʔá̰ɬ ja̰ː ḭštamaːkiːnḭ́ː ḭšluːwa̰ká̰n | ||||
| ta–la̰ʔ–a̰n–ɬ | ja̰ː | ḭš–ta–maːkíː–nḭː | ḭš–lúːwa̰–ka̰n | ||
| 3PL.SUB–ALTV–go–PFV | where | PAST–3PL.SUB–keep–PF | 3PO–snake–PL.PO | ||
| ‘they went to where they kept their snake’ | |||||
The first two valency-increasers are causatives. (5a) shows the cross-linguistically typical causative that adds a volitional agent to the event expressed by the verb; (5b) shows the stimulus morpheme ma̰ʔa-, which typically adds a non-human or inanimate causer, or is used to form verbs expressing caused internal or psychological states. (5c) illustrates the benefactive suffix,
-nḭ, which adds a beneficiary, maleficiary, recipient, or experiencer to an event. The instrumental applicative, liː-, shown in (5d), typically adds an instrument, but can also add motives. The comitative ta̰ː- seen in (5e) adds a co-actor that performs the action described by the verb along with the subject. The final applicative, the allative la̰ʔ-, adds a goal to a limited set of motion verbs, mostly those based on the verbs a̰n ‘go’ and min ‘come’.
RIGHT
Upper Necaxa Totonac has a total of six valency-increasing morphemes, two causatives and three applicatives. The causatives are maː- -niː ‘causative’ (6a) and ma̰ʔa- ‘stimulus’ (6b):
| (6) | a. | kimaːɬkaːkníːya̰ | |||
| kin–maː–ɬkaːk–niː–ya̰ | |||||
| 1OBJ–CS–hot–CS–IMPF:2SG.SUB | |||||
| ‘you are making me hot’ | |||||
| b. | ma̰ʔajikwánlḭ tsa̰má ʔawá̰ča̰ | ||||
| ma̰ʔa–jikwán–lḭ | tsa̰má | ʔawá̰ča̰ | |||
| STM–afraid–PFV | that | boy | |||
| ‘he frightened the boy’ | |||||
Of the two, maː- -niː is the cross-linguistically typical causative that adds a volitional agent to the event expressed by the verb; the stimulus morpheme ma̰ʔa-, on the other hand, typically adds a non-human or inanimate causer, or is used to form verbs expressing caused internal or psychological states.
The first of the four applicatives is the benefactive suffix, -nḭ shown in (7):
| (7) | na̰kšo̰ʔonḭyáːn ḭštapáɬ | |||
| na–ḭk–šo̰ʔó–nḭ–yaː–n | ḭš–tapáɬ | |||
| FUT–1SG.SUB–pay–BEN–IMPF–2OBJ | 3PO–price | |||
| ‘I’ll pay you its price’ | ||||
This suffix adds a beneficiary, maleficiary, recipient, or experiencer to an event.
The example in (8) illustrates the instrumental applicative, liː-:
| (8) | šuːnúːk naliːʔeːchiːkán ša̰ː | |||
| šuːnúːk | na–liː–ʔeː–chiː–kan | ša̰ː | ||
| bark | FUT–INST–back–tie–IDF | sweatlodge | ||
| ‘they will tie it onto the sweatlodge with bark’ | ||||
Not unexpectedly, this prefix typically adds an instrument to the event expressed by its base, but may also be used to add a motive for the event.
The comitative applicative adds a co-actor to the event:
| (9) | nakinta̰ːpína̰ | |||
| na–kin–ta̰ː–pin–a̰ | ||||
| FUT–1OBJ–CMT–go:2SUB–IMPF:2SG.SUB | ||||
| ‘you will go with me’ | ||||
As shown in (9), the co-actor performs the action described by the verb along with the subject.
The final applicative, the allative la̰ʔ-, adds a goal to an event:
| (10) | tala̰ʔá̰ɬ xa̰ː ḭštamaːkiːnḭ́ː ḭšluːwa̰ká̰n | |||
| ta–la̰ʔ–a̰n–ɬ | xa̰ː | ḭš–ta–maːkíː–nḭː | ḭš–lúːwa̰–ka̰n | |
| 3PL.SUB–ALTV–go–PFV | where | PAST–3PL.SUB–keep–PF | 3PO–snake–PL.PO | |
| ‘they went to where they kept their snake’ | ||||
The allative applicative combines with only a limited set of motion verbs, mostly those based on the verbs a̰n ‘go’ and min ‘come’.
Do not just give complex data and then expect the reader to figure them out
WRONG
Lushootseed oblique-centred relative clauses like that in (11) require nominalization of the embedded verb:
| (11) | x̌ʷul’ čəd ɬuləʔux̌ʷtxʷ tiʔəʔ ɬadsʔəɬtxʷ | ||||
| x̌ʷul’ | čəd | ɬu=lə=ʔux̌ʷ–txʷ | tiʔəʔ | ɬu=ad=s=ʔəɬ–txʷ | |
| only | 1SG.SUB | IRR=PROG=go–ECS | PROX | IRR=2SG.PO=NM=eat–ECS | |
| ‘I will just be taking [them] what you will feed [them] with’ | |||||
| (Hess 1998: 58, line 56) | |||||
As in all such relatives, the nominalized clause realizes its “subject” as a possessor.
RIGHT
Lushootseed oblique-centred relative clauses like that shown in the sentence in (12) require nominalization of the embedded verb, here ʔəɬtxʷ ‘feed sby with sth’:
| (12) | x̌ʷul’ čəd ɬuləʔux̌ʷtxʷ tiʔəʔ ɬadsʔəɬtxʷ | ||||
| x̌ʷul’ | čəd | ɬu=lə=ʔux̌ʷ–txʷ | tiʔəʔ | ɬu=ad=s=ʔəɬ–txʷ | |
| only | 1SG.SUB | IRR=PROG=go–ECS | PROX | IRR=2SG.PO=NM=eat–ECS | |
| ‘I will just be taking [them] what you will feed [them] with’ | |||||
| (Hess 1998: 58, line 56) | |||||
In this example, the headless relative clause tiʔəʔ ɬadsʔəɬtxʷ ‘what you will feed them with’ refers to the food, the oblique object of the verb ʔəɬtxʷ ‘feed sby with sth’; because the referent of the clause is an oblique object, the clause is nominalized with the nominalizing clitic s= and second person is realized as a possessive prefix, ad=, rather than as a subject clitic.
Do not simply end a section or a topic with a dataset
WRONG
A very common error writers make, usually in conjunction with the “don’t discuss data you haven’t presented” error, is giving data or placing an inset and then simply moving on to the next topic:
Datasets (and insets) should always be followed by a wrap-up either of the data (an explanation of the example) or of the topic
RIGHT
A very common error writers make, usually in conjunction with the “don’t discuss data you haven’t presented” error, is giving data or placing an inset and then simply moving on to the next topic:
Datasets (and insets) should always be followed by a wrap-up either of the data (an explanation of the example) or of the topic
Avoiding this error ensures good continuity and flow in your paper, and adhering to it usually forces you to think carefully about whether, indeed, you have made it clear enough to the reader what the data they just saw were actually about. It also gives you a chance to add corollary information that might be helpful to the discussion, such as the observation that following this last rule avoids the error of ending a paragraph with a dataset as well (since paragraphs usually enjoy some topical unity). Good stylistic practice usually dictates that the paragraph following a dataset not be indented (in styles with indentation on the first line of every paragraph).