n all accounts, Parmenides makes a marvelous argument in the
preters about how to read it. The only noncontroversial point in
interpreting the work seems to be that in it, Parmenides did something pro-
found to philosophy. Despite this collective obligation to acknowledge Par-
menides’ unique innovation (whatever it may be), it has become popular to
read Parmenides as relying on a modal fallacy to make his argument. This
would be an embarrassing mistake for such an influential work, especially
given the argument’s deductive appearance. In this paper, I will outline the
modal fallacy that Parmenides is accused of and argue for an interpretation
that is free of the fallacy.
The critical fragment we must examine to decide the question is
B6.1–2, in which the fallacy is supposed to occur:
crhV toV levgein te noei`n t
' ejoVn e!mmenai e!sti gaVr ei^nai,
As usual, Parmenides’ language admits of many permissible translations for
interpreters to quibble over, but this fragment is exceptionally difficult to
render. T. M. Robinson rightly points out that “the fragment is notoriously
replete with ambiguities” (627). Let us examine a few translations, then,
keeping in mind that classicists and linguists can marshal complex argu-
ments for and against each of them.
In Early Greek Philosophy, John Burnet translates the fragment in
It needs must be that what can be spoken and
thought is; for it is possible for it to be, and it is not
possible for what is nothing to be. (174)
In their commentary, Presocratic Philosophers, G. S. Kirk and J. E. Raven
follow Burnet and offer a similar translation:
That which can be spoken and thought needs must
be; for it is possible for it, but not for nothing, to
they were considered authoritative and widely influenced Presocratic schol-
arship. Although these translators did not explicitly accuse Parmenides of a
modal fallacy, one can detect it after formalizing their translations.
To show the fallacy, I will convert the key phrases from the fragment
into logical premises (note that the first premise comes from a previous
“the same thing exists for thinking and for being”
(Kirk and Raven 269).
(1) x can be spoken
↔ x can exist.
“it is possible for it [what can be thought and spo-
(2) What can be spoken
thought can possibly exist.
It is significant that when Schofield edited a second edition of Kirk and Raven’s collection, he
gave a new translation that avoids the fallacy altogether by leaving out the modal qualifiers: “What
is there to be said and thought must needs be: for it is there for being, but nothing is not” (247).
It is somewhat controversial what Parmenides has in mind here. Some consider it an unassuming
will see below how Robinson uses ascertainment to help avoid the fallacy.
It is very controversial how to construe “is there.” Whether Parmenides intended it to be a verb
paper, I have construed it existentially, but it may be productive to examine how each sense bears
on the fallacy. Mourelatos, in The Route of Parmenides, has argued that a speculative sense will avoid
the fallacy (60–63).
“it is not possible for what is nothing to be”
(3) What-is-not can’t possibly exist.
“what can be spoken and thought is”
∴ (C1) What can be spoken
thought exists. (2, 1 mo-
∴ (C2) Thus, what-is-not can’t be spoken
The alleged fallacy is to take premise (3) as a de re statement (i.e., P →
(P → P)). That what-
elors are unmarried men” is true. But the argument requires something
stronger to effect its conclusion. The scope of this modal operator (whether
it is a de re or a de dicto statement) means the difference between logical
tautology and metaphysical necessity.
If this is a correct translation and interpretation of B6.1–2, Par-
menides has certainly made a philosophical mistake. There are many con-
ceptions of possibility (e.g., logical possibility and nomological possibility),
and the Greek language remained ambiguous even between modal and
But we will need to consider only the broadest meaning
many things that are possible: it is possible that the earth has two moons
or that I ate spaghetti for dinner last night. Clearly, that these things are
possible does not imply that they are also actual. That something can be
does not imply that it obtains, at least not without additional argument.
However, if we read Parmenides as making this mistake, his argument fol-
lows smoothly except for its reliance on this fallacious premise.
The pieces needed to construct the fallacy have been agreed upon
some time to gain attention. G. E. L. Owen was the first to explicitly accuse
Parmenides of this modal fallacy, which he did in his watershed article “El-
eatic Questions.” Here, Owen follows Burnet’s translation of B6.1–2 with
some small stylistic differences:
What can be spoken and thought of must exist; for
it can exist, whereas nothing cannot. (94)
This translation carries the same logical troubles for Parmenides’ argument
that Burnet’s did. Owen describes Parmenides’ fallacy in a footnote:
See Kahn 722 and Robinson 633. e!sti can refer both to what is real and what is true.
This will become more clear when I treat Lewis’s interpretation below.
A, which can exist, is distinguished from B, which (poor
thing) cannot: invalid, for to say ‘nothing cannot exist’
is not to ascribe compulsory non-existence to anything
but to say that it is necessarily (truistically) true that what
doesn’t exist doesn’t exist, and this unexciting reformula-
tion disables the argument. The fallacy is the so-called de
re interpretation of modal statements. (94)
Owen’s criticism is identical to the criticism of premise (3) given above: B
cannot exist, but only in a de dicto formulation, not in a de re formulation,
and for Parmenides to rule it out as an object of thought or speech requires
the illegal de re statement.
Owen gives two reasons to believe that “this celebrated fallacy is the
point of the lines [B6.1–2]” (94). First, Owen believes this is the only trans-
lation that saves Parmenides from tautology, and the goddess clearly ex-
pects something more substantial than a tautology due to her injunction in
B6.5, “that is what I bid you consider,” and due to her allusions to the first,
wrong path in B2 (Kirk and Raven 270). Second, this translation fits nicely
with the point of B3 (which establishes premise (1) above).
Many scholars have followed Owen in ascribing this modal fallacy to
For example, Frank A. Lewis further presses the fallacy in his
appears that Parmenides presents some ways of inquiry and then argues by
elimination for one of them: the way of is remains, while the ways of is not
and of mixed (is and is not) fail. It is often supposed that mixed fails because
of the law of excluded middle: if either is or is not, then the conjunct of the
two (mixed) will be false. However, Lewis points out that this common read-
ing of Parmenides fails to preserve this argumentative structure of elimina-
tion; Parmenides’ modal upgrades to is and is not (added in B2.3–6) make
it impossible to eliminate mixed because of the law of excluded middle: “If
the two ways are represented by their modalized upgrades, the disjunction
of the two no longer makes an instance of the law of excluded middle” (2).
The contingent, modal statements “is but can not be; and is not, but can be”
are both still available, further ruining Parmenides’ argument by elimina-
tion (Lewis 2–3).
Hence, instead of relying on the law of excluded middle to eliminate
mixed, Lewis believes Parmenides’ denial of mixed relies on a modal fal-
lacy that he calls an “illicit modal shift”: “‘necessarily (is → is)’ becomes
‘is → necessarily is’” (5). This complaint is identical to the fallacy outlined
For example, in Philosophy Before Socrates, R. McKirahan quotes Owen on the modal fallacy and
enly cites Owen to substantiate his own claim that the argument “contains a subtle fallacy, treating
‘nothing’ as if it could intelligibly, even if falsely, be said to exist” (164).
in premise (3) above. Furthermore, Parmenides could deny the contingent
modal statements of mixed using this illicit modal shift. Hence, Lewis is
confident that the argument by elimination can be restored by simply sup-
posing that Parmenides was unaware of this fallacious modal shift, and the
rest appears cogent enough.
Owen and Lewis offer a tidy reading of Parmenides’ poem, but as-
suming a fallacy is a high price to pay, especially for interpreting an excep-
tionally deductive argument like Parmenides’. If one is reading charitably,
one ought to avoid readings that require such ugly fallacies from the outset.
Not all scholars attribute the fallacy to Parmenides. Still, it may be chal-
lenging to find a translation that is free of the fallacy and that can compete
with the otherwise graceful readings of Owen and Lewis. A non-fallacious
reading must also be philosophically and linguistically compelling.
Owen M. Goldin has given a possible escape for interpreting Par-
menides without a modal fallacy. His interpretation is intended to agree
with Owen’s “Eleatic Questions” with a single departure: Goldin denies
“Owen’s claim that Parmenides’ argument for the existence of any object
of reference or thought rests on fallacious modal logic” (19). Goldin points
out that in Owen’s version, the fallacy springs from the range of the modal
qualifiers, while in another version, which he argues is superior, the fallacy
results from the poem’s equivocal language.
According to this version, the
and existential without distinguishing them. In this case, the alleged fallacy
does not concern the scope of the modal qualifier, but involves an equivo-
cal shift from the potential meaning to the existential meaning. Because
Goldin cites convincing linguistic arguments to prefer his version, it is nec-
essary to reevaluate modality’s role in Parmenides’ argument.
Goldin points out that the philosophical ambiguity of Parmenides’
Parmenides’ argument is safe if his proposition amounts to “the principle
that what is thinkable is a possible being . . . [Indeed, in this case] we see
that Parmenides has explicitly considered the status of a possible existent;
its being is, as an availability, intention or goal” (27–28). If this was his posi-
tion, Parmenides is free to argue that all possible objects of thought already
Specifically, Goldin cites Kahn’s work on the ambiguity of the potential and locative (existential)
uses of ei^nai (Kahn argues that they were undistinguished in the Greek mind) and Cordero’s work
on toV or te in the manuscripts (he argues that it cannot be construed as Owen’s translation requires).
exist in some sense without committing any modal fallacy or equivocal
shift; interpreted this way, Parmenides is saying that thoughts are guaran-
teed existence by their nature, not by a modal argument from what can
exist. What is thought and spoken must exist qua thought or utterance.
Goldin’s reading is attractive, but I believe that it fails to preserve
large portions of Parmenides’ argument and that it relies on problematic
anachronisms. First, it is a vital element of the poem that the Way of Truth
is certain and in no way contingent. But if Parmenides considers thoughts
themselves to be real beings, he will be forced to embrace a completely
contingent reality that is at odds with his poem’s central theme. Suppose
thoughts are things for Parmenides, as Goldin proposes. These thoughts
either can be created by the mind or they cannot. If minds can actively cre-
ate, then the resulting aspect of existence is contingent (contra Parmenides’
salient point) on being thought by a mind. On the other hand, if minds
cannot create, it makes no sense to follow the goddess’s earnest command
to reason, nor does it make any sense to censure double-minded men as
the poem does. Some conceptions of mind may allow all thoughts to be
necessary products of the mind, but it seems impossible to consider Par-
menides’ argument in this way since his project presupposes possible mind
states from the outset. Although treating thoughts as entities is an interest-
ing ontological position that could be successful on its own and perhaps
even save Parmenides from a modal fallacy, it cannot exist peaceably within
Furthermore, Goldin’s reading of Parmenides as a proto-Meinong
is historically implausible. Although Goldin delineates some fine distinc-
tions between Meinong and Parmenides in his article, it appears prima
facie impossible to argue that the Greeks at the time of Parmenides made
such distinctions. One would expect massive reactions to such an interest-
ing position, yet no ancient philosopher shows any signs of appreciating
it. On the other interpretations covered, one can at least find reactions, or
even misreadings, of Parmenides in the works of Melissus or the Atomists.
If Goldin’s interpretation is correct, then Parmenides would have gone
interpretation does not cohere with the history of thought any better than
it fits within Parmenides’ own premises.
It may seem odd, or even circular, to use misreadings as evidence for my preferred interpretation
influence their successors deliberately and accidentally, but in both cases their influence is a func-
tion of attention. A misreading can be especially revealing if one can pinpoint its cause. Correct
understandings make for straightforward evidence, but misreadings can reveal more about the
philosophical milieu and the priorities of contemporary readers. Modern interpreters should pay
attention to both in order to construct a thorough story of developing ideas.
Perhaps Parmenides’ argument can benefit from an alternative trans-
lation of B6.1–2. In his article, “Parmenides on Ascertainment of the Real,”
T. M. Robinson gives various translations of key fragments and defends
them linguistically. Consider his translation of B6.1–2:
Necessarily, what is there to pick out and ascertain
is real for it is there to be real, whereas nothing is
Notice that in this translation the pronoun “it” no longer refers to
“what can be spoken or thought of ” as it did on previous translations.
Now “it” refers to “what is there to pick out and ascertain.” This is a major
shift away from Burnet’s translation. An external world (what is there) has
supplanted an internal one (products of thought and speech), effectively
reversing the direction of the argument. On this translation, Parmenides
does not argue that our thoughts constitute reality, but rather that reality
supervenes on our thoughts (at least, on our thoughts about ascertaining
the world). This is important because “Parmenides is at once absolved of
the elementary modal fallacy to which the translations of Burnet, [Kirk and
Raven], and Owen appear to commit him” (Robinson 628).
Consequently, Robinson believes it is impossible to read Parmenides
as holding that thinking and the real are identical. Instead, “[Parmenides]
argued for the epistemologically unimpeachable view that there is a neces-
and in doing so laid firm foundations for the entire future science of episte-
mology” (Robinson 632). Moreover, the meaning of Parmenides’ conclusion
is also that, necessarily, real things can be referred to. Although modal and
moral claims remain grammatically joined in the poem, Parmenides has ap-
proached the topic with sure first steps under this interpretation. Further,
his translation of B6.1–2 indicates that Parmenides “is discussing availabil-
ity, not possibility,” which further discredits accusations of modal fallacies
(Robinson 633). None of these advancements are traditionally credited to
Parmenides, but Robinson has shown that they are available in the text.
In considering interpretations without modal fallacies, I have shown
how Goldin’s interpretation failed both within Parmenides’ argument and
within its historical context. Robinson’s translation of B6.1–2, on the other
hand, offers interesting insights on both fronts. First, within Parmenides’
own philosophy, it becomes clear that he is not equating thinking with real-
ity. Rather, Parmenides argues that what is real can necessarily be referred to.
This position becomes especially intriguing for the remainder of Parmenides’
philosophy, especially his Way of Opinion. For example, consider the sta-
tus of Parmenides’ cosmology, often treated as the exemplification of the
Way of Opinion. With Robinson’s translation, the contingent objects of
the cosmology have some virtues. Although they remain contingent (unlike
the necessary propositions in the Way of Truth), they can at least be reliably
ascertained (thanks to the propositions deduced in the Way of Truth, which
under this interpretation show that what we ascertain is real). Hence, while
there is a difference between knowing what must be and knowing what is
and may not be, on Robinson’s reading the second type of knowledge is not
worthless, as it is on other readings. Moreover, this position helps us make
sense of the goddess’s demand to learn the Way of Opinion and account
for the remarkably useful astronomy that Parmenides discovered. Interpreta-
tions that completely deny the usefulness of the Way of Opinion fail to ap-
preciate one of the greatest scientific observers of an era.
Second, this reading also illuminates Parmenides’ historical influ-
ence. Robinson shows that under this interpretation, Parmenides’ argu-
ment gives a structural precedent for Plato’s Timaeus 28c–29a (Robinson
628). Here, Plato employs an argument that is analogous in structure to
that of Parmenides B6.1–2. Plato examines a disjunction between created
and uncreated paradigms of cosmogony. This is similar to the disjunction
examined by Parmenides: what can exist is comparable to the uncreated
paradigm, and what cannot exist is comparable to the created paradigm. In
both instances the second item is ruled out: what is not could not be, and
the cosmos could not be created, given the nature of the forms. Neither
is possible, yet they are used as counterbalances to affirm their comple-
ments. Again, we see Parmenides influencing Plato in Republic 476e–480
(Robinson 632). Here, Parmenides’ emphasis on the real and what can be
ascertained appears as Plato discusses what constitutes knowledge, and he
does so in strikingly Parmenidean terms. Again, the upshot is analogous:
for Parmenides, this premise supports the One, while for Plato it supports
his Forms. This interpretation, then, fits the historical picture well.
In conclusion, it appears that Parmenides’ argument doesn’t need
a modal fallacy to remain (otherwise) cogent and compelling. The above
interpretations are strikingly divergent (which is perhaps characteristic
of interpretations of Parmenides in general). Lewis can be thought of as
an extension of Owen with a more complete account of Parmenides’ argu-
ment by elimination, but both interpretations rely on the modal fallacy to
interpret Parmenides. Goldin’s Parmenides builds his reality from objects
of thought, whereas Robinson’s Parmenides builds his objects of thought
from reality, and both avoid the modal fallacy in reading Parmenides. Yet
Goldin fails to preserve Parmenides’ larger argument and can’t establish
reasonable historical reactions to his interpretation. If we agree with Rob-
inson’s translation of B6.1–2, I have shown how one can maintain Par-
menides’ wider philosophy, particularly his cosmology, as well as his po-
sition in the history of thought, as seen in works of Plato. This reading
of Parmenides is as graceful as Owen’s or Lewis’s, and it comes with the
significant bonus of avoiding the fallacy.
Burnet, John. Early Greek Philosophy. 4th ed. New York: Meridian Books,
Cordero, Nestor Luis. “Les Deux Chemins de Parménide dans les
————. “L’Histoire du Texte de Parménide.” Études sur Parmenide. Ed. P.
Abenque. Paris: J. Vrin, 1987. 3–24.
Graham, Daniel W. The Texts of Early Greek Philosophy. Cambridge:
Cambridge UP, 2010.
Goldin, Owen M. “Parmenides on Possibility and Thought.” Apeiron: A
Kahn, Charles. “The Thesis of Parmenides.” Review of Metaphysics 22.4
Kirk, G. S. and J. E. Raven. The Presocratic Philosophers. Cambridge:
Cambridge UP, 1957.
Kirk, G. S., J. E. Raven, and M. Schofield. The Presocratic Philosophers. 2nd
ed. Cambridge: Cambridge UP, 1983.
Lewis, Frank A. “Parmenides’ Modal Fallacy.” Phronesis 54.1 (2009): 1–8.
McKirahan, Richard D. Philosophy Before Socrates. Indianapolis: Hackett,
Tugwell, Simon. “The Way of Truth.” Classical Quarterly 14.1 (1964): 36–41.
Curd, Patricia. The Legacy of Parmenides. Princeton: Princeton UP, 1998.
Kahn, Charles. “The Greek Verb ‘To Be’ and the Concept of Being.”
Foundations of Language 2 (1966): 245–65.
————. The Verb Be in Ancient Greek. Indianapolis: Hackett, 2003.
Mourelatos, Alexander. “Pre-Socratic Origins of the Principle that There
are No Origins from Nothing.” Journal of Philosophy 78.11 (1981):
O’Brien, D. “Le Poème de Parménide: Texte, Traduction et Essai Critique.”
Palmer, John. Parmenides and Presocratic Philosophy. New York: Oxford UP,
Robinson, T. M. “Parmenides on the Real in its Totality.” Monist 62.1
Stannard, Jerry. “Parmenidean Logic.” Philosophical Review 69.4 (1960):