Borg’s Minimalism

Borg takes minimalism to be defined by the following four claims[CITATION Bor2a \n \t \l 2057 ]:

i. Well-formed declarative sentences express truth-evaluable content.

The idea embedded in the claims (i) and (ii) is that semantic content of sentences attaches to
formally specified objects and that the route to meaning travels exclusively through the syntax and
the lexicon. That is to say that the syntactic structure and the lexical content attached to words
provides the sufficient and complete input to a semantic theory, which would then yield an account
of the meaning of all well-formed sentences. what supports these claims is that linguistic
competences such as productivity and systematicity can only be explained if we assume that
semantic content is delivered through recursively specified rules operating over syntactically
described objects.

So, why might we think that purely lexico-syntactically described objects (sentences) are (at least
sometimes) capable of expressing truth-evaluable, propositional contents?

1- It is extremely natural for competent speakers to treat sentences like grass is green and
l'herbe est verte as conveying information which represents the world as being in a certain
way and is capable of being true or false. It is also because that we allow multiple sentences
(some in different languages), to express the same content and therby make the same claim
about the world that we consider positing a propositional content which these sentences
share.
2- The fact that valid inferences can be made about shared contents when faced with
sentences containing belief ascriptions. For example, we can infer from the sentences like
‘Jack believes Germany won the war’ and ‘Steve believes Germany won the war’ that there
is something both Jack and Steve believe. In addition to that, if we can also judge the truth
or falsity of their believes. And this kind of practice implies that the sentences in the scope of
‘believes’ express a content that is truth-evaluable and attributable to the subjects.
3- It is an obvious fact that arguments expressed through natural language sentences can be
judged as valid or invalid, or that a sentence can contradict anther but this is only possible if
we assume that natural language sentences are capable of expressing propositional content.

my claim is that our default position is that at least some sentences do express truthevaluable,
propositional content.

recovery of semantic content should be possible without appeal to current speaker intentions.
 That is to say, when we talk of semantic content as being delivered for sentences relativized to
contexts of utterance (as we must in order to accommodate overt context- sensitivity) the kinds of
contexts in play will be formal given by ordered sets of objective parameters (like time and world of
utterance) which do not include the intentional states of the speaker. The argument for this
constraint is two-fold:

1-such a constraint is definitive character of formal approaches to meaning. The semantic content of
sentences must be determined by lexico-syntactic features and be formally tractable. but allowing
speaker intentions to be necessary in recovering the semantic content (atleast for some expressions)
runs counter to the ethos of formal semantics. Also if we did allow speaker intentions to be
semantically relevant In some cases, why might we use them in all cases? Therby allowing semantic
to come close to the communicated content. “That is to say, if we allow that I need to know what
you are thinking in order to recover the semantic contribution of an utterance of ‘that’ (because I
need to work out to which object you intend to refer), why shouldn’t it also be the case that I need
to know what you are thinking in order to recover the semantic contribution of, say, ‘red’ as well
(because I need to know if you intend to mean red on the inside, red on the outside, etc”.

2-the second reason concerns the relationship between minimalism and the modular theory of
mind. A strong reason that motivates us to endorse the idea that semantic content must be
amenable to treatment via a formal,syntax-driven theory is that only in this way could our semantic
competence be underpinned by a genuine Fodorian module. where what is essential to a Fodorian
module is that it be a computational device, a Turing machine. Yet, if we want to tell a Fodorian,
modular story about our semantic competence then, on the one hand, semantic content must be
recoverable via purely lexico-syntactic means and, on the other, the processes operating over those
syntactically described items must themselves be formal, computationally tractable operations.

The problem then is that, while Fodorian modules are computational, syntax-driven devices,
mindreading is an abductive, syntax independent process par excellence. For instance, to work out
if, in moving his cricket bat thus and so, Flintoff is preparing for a defensive shot, or to cut the ball,
or perhaps even to whack the fielder at silly mid-off , one needs to know a wealth of information
about the situation, including information about what Flintoff wants or believes. Furthermore,
learning something new about the situation can always change our views on the best interpretation
of Flintoff ’s actions (background information about Flintoff ’s batting style may lead me to believe
he is intending an over-ambitious slog, but learning what the fi elder at silly mid-off has just said
may cause me to revise this judgement). The reasoning involved in the attribution of mental states is
thus, it seems, inductive, defeasible and content driven.

properly formal semantics must be one which steers clear of the ‘magic of pragmatics’ From this
perspective, a theory is free of the magic of pragmatics if it offers an account of our semantic
competence which is computationally tractable. That is to say, if it maintains that semantic content
can be recovered on the basis of limited, encapsulated information about lexical content and
grammatical structure, rather than requiring appeal to, potentially, anything a subject believes.

So, my variety of minimalism will see the semantic contribution of context as doubly constrained: on
the one hand, all versions of minimalism agree that context contributes to content only when
syntactically triggered and hold that the expressions which can perform this kind of triggering are
pretty limited. In addition to this, however, my more stringent version of minimalism will also hold
that what context is able to supply is a set of objective parameters (such as world and time) and not
rich, intensional aspects of the context of utterance, such as speaker beliefs or intentions. It is this
doubly constrained version of minimalism which I’ll seek to defend in the rest of this essay.
Reasons for endorsing minamlism:

1-The intuitive appeal of minimalism. Pretheoretically the idea that sentences convey information
which is capable of being true or false, and that the information a given sentence conveys is simply a
function of, more or less, the words it contains and the way they are put together, seems pretty
appealing. Now, as we are going to see time and again in what follows, adopting this kind of
simplistic approach to semantic content leads the minimalist to make predictions about semantic
content which diverge from the attributions of content which ordinary speakers make to the things
they say in conversational exchanges So, for instance, minimalism claims that the sentence ‘There is
nothing to eat’ means that there is nothing to eat (in some universal domain) . Yet of course what
someone conveys when they utter this sentence is likely to be a much narrower content like there is
nothing to eat in the fridge or there is nothing appealing to eat in the kitchen . Furthermore, this
point is entirely general: take any sentence of natural language and minimalism is likely to predict a
literal meaning for it which is at odds with the content likely to be conveyed by an utterance of that
sentence.

So why do we need an account of semantic content that departs from our intuituions about what is
said , the question is, would anyone want an account of semantic content which departs from what
is said conversational exchanges?.

1-, such an approach does fit with at least some pretheoretical intuitions

, I think it is not right to portray minimalism as running counter to all pre-theoretical intuitions about
utterance-level content. Given the right context (i.e. one where subjects are asked to reflect on
‘literal’ or ‘strict’ meaning) ordinary interlocutors can and do grasp exactly the kinds of contents
minimalism predicts. Children, legislators, and philosophers often seem to operate precisely with the
kind of strict, syntax-based model of meaning which minimalism suggests, so to claim that the
readings the account predicts fail to match any of our intuitions about utterance content would be
far too strong.

2- the idea that semantics should in general capture intuitions about speech act content is ill-
conceived

it is simply a mistake to require a semantic theory to be answerable to intuitions about speech act
content, for semantic content is one kind of thing (a repeatable, codifiable, rule-governed kind of
thing) while speech act content is another kind of thing altogether (a potentially unrepeatable,
nebulous, context-governed kind of thing). The sentence ‘It’s raining’ can convey that it is raining
here, raining there, that it is my favourite weather in Palo Alto today, or any of an indefinite number
of things beside. Though clearly there must be a connection between the meanings sentences have
and the things we can say using them—roughly, it must be in some way because sentence s means
p that we can use it to say q —still there is simply no reason to believe we can isolate someone (or
just a few) of the things that can be conveyed by a sentence and treat them as yielding the literal
content for that sentence.

3-if we jettison the lexico-syntactic borders on semantic content we will be left unable to reconstruct
a semantics/pragmatics divide at all.

The third reason for accepting minimalism is that moves away from an exclusively lexico-syntactic
route to semantic content (which seem to be necessary if we want a semantic theory that fits better
with speaker intentions) would leave us all at sea in pragmatic space. the claim is that if we dissolve
the syntactic walls on what counts as semantic content we will be left with no way to reconstruct
any walls at all. That is to say, we will lose all distinction between literal meaning and speech act
content. If we open the door to let a little bit of (not-syntactically demanded) pragmatic content
inside the citadel of semantics we will at once be overwhelmed by a tsunami of pragmatic content,
eroding entirely the semantics/pragmatics distinction. Contrary to some claims, such an approach
does fi t with at least some pretheoretic intuitions, second the idea that semantics should in general
capture intuitions about speech act content is ill-conceived and third, if we jettison the lexico-
syntactic borders on semantic content we will be left unable to reconstruct a semantics/pragmatics
divide at all. Finally, as argued in Borg 2004a , it seems that only an intention-insensitive semantics
like minimalism could fi t within a computational, Fodorian module for linguistic comprehension and
production.

4-only an intention-insensitive semantics like minimalism could fi t within a computational, Fodorian

minimalism is the kind of semantic theory which stands a chance of fitting within a modular story of
our linguistic competence and there is independent reason to think a modular story about our
linguistic competence is right. However, minimalism fits within a Fodorian module only on the
condition that it involves deductive reasoning processes alone. So minimalism, as I construe it,
should be constrained to account for linguistic understanding in terms of discrete, deductive, syntax-
driven reasoning processes, rather than via the murky world of content-driven inference to the best
explanation.

Semantic relativism:

MacFarlane 2007argues that we need to adopt full-blown assessor-relativism (whereby the truth-
value of the content a sentence expresses is relativized not merely to additional features in the
context of utterance, but also to features within the context of assessment, so that we lose entirely
the notion of absolute truth for linguistic contents), but also to maintain that there are some
contexts where if A accepts p and B rejects p, then it cannot be that both A and B are accurate in
their judgements. 39Whether this position is defensible or not, the point to notice here is that, once
we start exploring what could motivate the relativist view as against indexicalism/contextualism, it
seems that the motivation speaks in favour of the most radical version of the view (i.e. the idea of
assessor-relativity); yet we might also ask, with Wright ( 2008 , forthcoming) and others, whether
such evidence about our linguistic practices can really be strong enough to motivate a truly radical
theory like assessor-relativism. 40