lemÌc. Lemizh grammar and dictionary

A sketch of pragmatics III. Discourse

Communication is first and foremost a kind of constrained mind-reading with linguistic … codes just providing evidence (often rich and detailed evidence, but never a complete encoding, never a proof) of the thoughts being communicated.

(Robyn Carston. Thoughts and Utterances)

Utterance modifiers

Utterance modifiers (UMs) are words and phrases that comment on the act of uttering a sentence, such as ‘frankly, confidentially, in case you’re interested’. They typically occur at the beginning of the sentence in question, being followed by a comma: Frankly, I like tortoises. means ‘Speaking frankly, I like tortoises’ or ‘I am saying frankly that I like tortoises’. A more elaborate UM is As you like tortoises, there are some on the shore. which conveys that I am saying this because you like tortoises (and therefore you might be interested in this piece of information).

They differ from sentence adverbs such as ‘luckily, unfortunately, naturally, ironically, possibly’, which comment on the content of an utterance: Luckily, Zarathustra will speak with us. cannot be paraphrased as ‘Speaking luckily / I am saying luckily that Zarathustra will speak with us’, but rather means ‘It is / We are lucky that Zarathustra will speak with us’.

Here are some examples. More can be found in Bach (1999), pages 356–358, from where most of the categorisation is taken:

CategoryUtterance modifiersApproximate meaning
(U1)veracitivesfrankly, indeed, believe it or notasserting the truthfulness of the utterance
(U2)emphaticsmark my words, it cannot be overemphasised thatemphasising the utterance
(U3)mitigativesneedless to say, it’s none of my business butmitigating the utterance
(U4)formulationalsmetaphorically speaking, without exaggerationcharacterising the phrasing of the utterance
(U5)secretivesconfidentially, between you and me, off the recordmarking the utterance as secret/confidential
(U6)scopingslogically/morally/botanically (speaking), in the kitchenrestricting the scope of the utterance
(U7)relationalsspeaking as your friend/doctor/fatherappealing to the relation between speaker and addressee
(U8)explanatoriesin case you’re interested, just so you know, as you like tortoisesexplaining the reason for the utterance

Many UMs comment on a relation to previous utterances, such as ‘furthermore’, or sometimes to following ones, such as ‘first of all’. These are called discourse connectives. Again, some examples:

CategoryDiscourse connectivesApproximate meaning
(D1)positionalsfirstly, first of all, finallyindicating the position in an enumeration
(D2)additivesfurthermore, besides, moreoveradding related information
(D3)conclusivesto sum up, all in allsumming up what has been said
(D4)formulationals*in other words, metaphorically speaking, in a nutshellreformulating an utterance (typically for clarification)
(D5)topicalsanyway, to change the subjectchanging the subject to something seen as more important by the speaker
incidentally, by the wayinserting a parenthetical statement not related or not central to the current subject
to get back to the point/subject, squabbles asidereturning to a previous subject, seen as more important by the speaker
wellreluctantly saying something not sufficiently relevant to the current subject (‘relevance hedging’)
(D6)illustratives†for example, for instancegiving an example
(D7)contrastives†however, yet, on the other handcontrasting new information to what has been said
(D8)concessives†nevertheless, stillsaying that the following is true despite what has been said

* (U4) and (D4) overlap: both refer to the utterance’s phrasing; the former (so to speak) in isolation, the latter in relation to what has been said previously.
† On the question of whether these are UMs at all see below.

Discourse connectives in Bach (1999) are listed in types 1–7; numbers 8, 9, 13 and 14 also contain some examples. However, we won’t analyse every single one of them.

Words and phrases used as UMs often have other, non-pragmatic, functions as well. In such uses, they are not necessarily sentence-initial. Thus, the highlighted portions of He spoke frankly about tortoises. As you like tortoises, I’ll show you some. There are no tortoises on the shore, but let’s go there anyway. are parts of the explicatures as opposed to commenting on what is said.


UMs bear some similarity to CIs (and have sometimes been classified as such) because they convey information additional to the basic explicature, are encoded and therefore not defeasible, and are detachable by omission. However, they fall inside the scope of subordinate clauses, and so neither project nor are speaker-oriented: I’ll be able to read Jacopo’s texts if, in a nutshell, I find the password. She dances so gracefully because, first of all, she practices several hours a day. Achilles told me frankly that he liked tortoises. Example (9) attributes the frankness to Achilles, not the speaker. However, it does not say that Achilles characterised his utterance as frank himself. This is yet another difference to CIs: UMs cannot comfortably be reported in indirect speech.

In Lemizh

Pronoun referring to the parole (U1U8, D1D4)

A comment on the act of uttering a sentence is, in Lemizh terms, a comment on the parole. We achieve this by referring to the parole with the relative pronoun wà. as the main predicate. This works for ordinary UMs as well as for some discourse connectives (which sometimes use the pronoun fà. to refer to the previous parole).

wà ywbà ràhy fkrÌjy.The parole is frank.Frankly, I like tortoises. (U1)
PIn−1-fact1 frank-acc-fact2 like-fact-acc2 tortoise-acc-acc3.
wà krìjan {saní} ráhy fkryjè zèi.This parole [and the previous one] form an ensemble.Furthermore, tortoises like me. (D2)
PIn−1-fact1 ensemble-dat-partfact2 {PIIn−3-partfact-dat3} like-fact-acc2 tortoise-acc-nom3a PIn−3-nom-dat3.
wà zvecè swnatRakskày fkrÌje.The sender of the parole is a friend.Speaking as a friend, don’t believe tortoises. (U7)
PIn−1-fact1 friend-nom-nom2 believe-fact-should-fact-opposition-fact-acc2 tortoise-acc-nom3.
màxky.The content of the previous parole is that they lie.In other words, they lie. (D4)
PIIn−1-fact1 lie-fact-acc2.

Many UMs are phrased as factive brackets parallel to ‘frankly’, such as:

The discourse connectives ‘metaphorically’ and ‘to sum up’ have an implicit pronoun referring to the previous parole: wà wrOnÌa {saí} —. ‘This parole-acc is a metaphor for the previous one-dat; the previous parole is expressed metaphorically with this one’. It hardy ever needs to be overt, though.

As this approach marks UMs explicitly as such, we have to distinguish them carefully from ordinary subordinate clauses.

là ráhOl ziè fkryjý dmáta zeèl fÌen.You liking tortoises is the reason for me showing you some. (ordinary subordinate clause; desorption to get the persuasive clause to the front)As you like tortoises, I’ll show you some.
do-fact1 like-fact-psu2 PIn−3-dat-nom3a tortoise-acc-acc3 see-fact-fact2 PIn−3-nom-caus3a PIIn−1-acc-partnom3.
wà ráhOl ziè fkryjý mìly fyìn splÌjor.You liking tortoises is the reason for me talking. (utterance modifier)As you like tortoises, there are some on the shore. (U8)
PIn−1-fact1 like-fact-psu2 PIn−3-dat-nom3a tortoise-acc-acc3 make-cons-acc2 PIIn−1-acc-partdat3 shore-acc-sce3.

Likewise, ‘If I may interrupt, I’ll tell you something important’ is an ordinary conditional clause, while ‘If I may interrupt, Achilles and the tortoise have just been kidnapped’ is a mitigative utterance modifier (and probably a discourse modifier). The same distinction has to be made between ‘Have you seen the tortoise again?’ and ‘Again, have you seen the tortoise?’.

Verbs of communication (D5)

Lemizh uses specialised verbs of communication to comment on subject changes and the like. They take the actual utterance as their accusative object. Common words are:

ycvà.changesubjectsay something-acc,changing the subject to something seen as more important by the speaker
Ràjd.digressinserting a parenthetical statement not related or not central to the current subject
bàx.backtopointreturning to a previous subject, seen as more important by the speaker
wellreluctantly saying something not sufficiently relevant to the current subject (‘relevance hedging’)

For the sake of consistency with other UMs, these verbs can be interpreted as having been absorbed by the pronoun wà..

wà Rajdà lapày dmìlty ksmÌse tÌy.
PIn−1-fact1 digress-fact-fact2 do-fact-ask-fact-acc2 see-cons-acc3 squirrel-acc-nom4 this-acc-acc5.
Ràjd lapày dmìlty ksmÌse tÌy.Incidentally, have you seen that squirrel?
digress-fact1 do-fact-ask-fact-acc2 see-cons-acc3 squirrel-acc-nom4 this-acc-acc5.
bàx wdriRgà làxty jàxy splÌji.Returning to the previous subject is the end of fighting.Squabbles aside, I’d like to go to the shore.
backtopoint-fact1 fight-egr-fact2 want-fact-acc2 move-fact-acc3 shore-acc-dat4.
These verbs also comment on indirect speech:
Rájd axileysè veì dmìlty ksmÌse.Digressing, Achilles told me that he had seen a squirrel.
digress-fact1 Achilles-acc-nom2a PIn−2-nom-dat2 see-cons-acc2 squirrel-acc-nom3.

Non–utterance modifiers (D6D8)

At least from a Lemizh view, illustratives (‘for example’), contrastives (‘however’) and concessives (‘nevertheless’) aren’t utterance modifiers as they comment on previous information, i.e. the content of a previous parole (the main predicate and its objects) as opposed to that parole itself. I won’t judge about their status in English.

làcw veì dmÌa. à hànca telmÌxi.The action of cleaning the bicycle is from the set of help he gave me. (Pronoun refers to main predicate of previous sentence.)He helped me a lot. For example, he cleaned my bycicle. (D6)
help-fact1 PIn−2-nom-dat2 3/4-acc-fact2. PIIn-fact1 clean-partfact-fact2 bicycle-acc-dat3.
avnatmà proxÌi.(tmÌ. ‘but’ marks unexpectedness based on previous information)However, he didn’t feed my badger. (D7)
eat-fact-not-fact-but-fact1 badger-acc-dat2.
rahdmáj proxyè fanàOlm.The badger rather liked him despite that. (Pronoun refers to main predicate of previous sentence.)Nevertheless, the badger rather liked him. (D8)
like-fact-5/8-fact1 badger-acc-nom2a PIIn−1-fact-not-fact-qualpsu2.

In the last example, we could move the Lemizh equivalent of ‘nevertheless’ to the front by a desorption.

There is a large number of constructions that connect utterances in a discourse but aren’t utterance modifiers; see for example ‘actually’ and ‘after all’ in the dictionary.

Indirect speech

In contrast to English, UMs can be effortlessly included in indirect speech; the pronoun, if present, will automatically point to the inquit.

wáx axileysè wày ywbà ràhy fkrÌjy.Achilles’s speaking is frank.Achilles said that, frankly, he liked tortoises. (awkward)
Achilles said, ‘Frankly, I like tortoises’.
speak-fact1 Achilles-acc-nom2a PIn−1-fact-acc2 frank-acc-fact3 like-fact-acc3 tortoise-acc-acc4.
wáx axileysè Ràjdy dmìlty ksmÌse.Achilles said, ‘Incidentally, I have seen a squirrel’.
speak-fact1 Achilles-acc-nom2a digress-fact-acc2 see-cons-acc3 squirrel-acc-nom4.

Note that, in the second example, ‘speak’ (Achilles’s parole) cannot absorb ‘digress/incidentally’ because the latter has an outer accusative, not a factive.

Use utterance modifiers sparingly: recall that everything clear from context should be omitted.

Conversation structure in a nutshell

Conversation among two or more people is structured by a system of turn-taking: each turn consists of one or more pragmatically self-contained elements, the turn construction units (TCUs), which may be single words (‘Yes’), phrases (‘Above the moon’), sentences, or possibly whole stories. The end of a TCU, called a transition relevance place (TRP), can usually be prediced very reliably by listeners by a variety of verbal and nonverbal cues, allowing for a minimum amount of pauses, overlaps and unintended interruptions. Cues include interjections, intonation, eye contact, gestures and posture.

Reaching a TRP, the speaker may

We will only mention three kinds of cues here: hesitation markers (‘er, uh, um’) indicate that the speaker is pausing to think and the pause is not a transition relevance place. Backchannels (‘uh-huh, yeah, right, oh’, various facial expressions) are reactions by the listener that do not interrupt the current TCU but signify attention, understanding, and often agreement, surprise, amusement, anger, or other emotions. A turn request (‘um, wait’) is an attempt to gain the floor at a non-TRP.

Much could be said on adjacency pairs (co-occurring pairs of turns such as greeting+greeting, question+answer, request+fulfillment/refusal), but nothing that is specific to Lemizh; so we will refer our readers to Levinson’s Pragmatics.

While details such as employed cues and their meaning, average pause length and amount of overlap vary to some extent across languages and cultures, the basic principles have been found to be essentialy universal. Umberto Eco, however, has pointed out that this theory never caught on among Italian linguists, who find it natural that everybody is talking at the same time, and this works just fine.

In Lemizh

The Lemizh, like most other cultures, take turns in conversation: the Italian model would make it impossible to use type II pronouns to refer to the previous parole.

The most common hesitation marker is the pronoun wà., lit. ‘this parole’, more informally translated as ‘I’m saying …’. It also serves as a low-key turn request. The Lemizh are restrained at backchannelling: as we know, everything clear from context should be omitted. The pronoun fÌ. ‘yes’ signals agreement, nà. ‘no’ dissent. Emotional reactions are typically facial expressions and non-grammatical vocalisations, as they cause less interference with the (grammatical) sentences the speaker is producing.

Residual issues

Speech act theory is not treated on these pages. Its concept of illocutionary acts more or less corresponds to propositional attitudes such as believing in the truth of one’s statement, feeling sorry for someone, expressing a (speaker’s or addressee’s) desire, or promising something. Perlocutionary acts are effects of what is said on the addressee, such as convincing, persuading, eliciting an answer, reassuring, scaring, etc.; they are largely language-independent, and to some extent dependent on culture, and at any rate beyond the scope of this website. Indirect speech acts are conveyed by implicatures such as ‘Susan believes that her kiwis were too sour. I want you to ring Susan and cheer her up’. See also the paragraph on speech acts in unit 9 of the tutorial. More on speech act theory can be found in the literature given below, including Sperber and Wilson (1995).

I am planning to add a chapter on politeness some time in the future. Finally, while some remarks on style and registers can be found in the dictionary, a coherent account will probably remain forever beyond the scope of this website.